LTE Protocol Stack Physical Layer

LTE Development, Conformance Test, Optimization
Certification Course Amateur Level (3PCA-L1)
LTE
Physical Layer
LTE Protocol Stack
Author: Surya Patar Munda
3PCA-L1
LTE Protocol Stack- 1
surya.patar@3gnets.in
Preface:
Dedication This book is dedicated to my family who has given me support to complete this book.
The colleagues in office have given me encouragement to start and complete this book. My hearty
thanks to all of you. The first release is printed with many terms unexplained and even sentences are
shortened but intended to cover in this book. They will gradually be expanded in next release. Please
do write me on the email given in the pages below to improve.
Who is this book for?

Over the years I have seen the telecom industry struggling to get right people with sufficient domain
knowledge in 2G or 3G or 4G. The specification is very huge and it is often horrendous to go through
the details. This book is referring most of the time with respect to LTE 3GPP specification, Rel-10.
This is an effort to consolidate information in an organised way to give a methodical way of
understanding LTE. This is a very good start for an engineer who is either going to pursue:
LTE Protocol Stack Development

LTE ConformanceTesting
LTE Network/RF Optimization
LTE entities (UE and Network both) troubleshooting
If you need 3GNets LTE Physical Layer for Amateur Level (3PCA-L1), you need this course. This
knowledge and level is required for the next level Professional Level (3PCP-L1) where you can
be trained for higher level with Hands on Projects and real implementation. Full Amateur level
courses are:
LTE Physical Layer LTE L2 Layer - MAC, RLC, PDCP LTE RRC
LTE NAS
(3PCA-L1)
(3PCA-L2)
(3PCA-RRC)
(3PCA-NAS)
About Author:
Surya Patar Munda has been in Telecommunications Since 1987 and has gone through the life cycle
of Software Development, Software Testing, Network Deployments, Integration, Testing,
Troubleshooting, Handphone Testing with Specification etc.. a full round of the Telecom industry. He
has worked with Motorola, Nortel Networks, Spirent Communications, Sasken etc. companies with full
round cycle. The Software engineers midset and Testing engineers mindsets are different and so is
the mindset of an RF optimization engineer. This book will cater to all.
Author also conducted many trainings for Telecom industry and has a very good understanding of
what kind of requirement is there for engineers. The goal is not just what and how does it work, but
also the goal is how do I start implementing and how do I test.
Edition: July 2013

Price: Rs.299
LTE Physical Layer- 3PCA-L1 Certification
email: suryapatar@yahoo.com
Page 2
Contents
1.
Downlink Physical Layer ................................................................................................................. 7

1.1
OFDMA Principles ................................................................................................................... 7
1.1.1
OFDM - Orthogonal Multiplexing Principle ...................................................................... 7
1.1.2
Peak-to-Average Power Ratio and Sensitivity ................................................................ 9
1.1.3
Timing Offset and Cyclic Prefix Dimensioning ................................................................ 9
1.1.4
OFDMA Parameter Dimensioning .............................................................................. 10
1.1.5
Physical Layer Parameters for LTE .............................................................................. 10
1.1.6
Transmission Resource Structure ................................................................................. 11
1.2
Synchronization and Cell Search .......................................................................................... 15
1.2.1.
Synchronization Sequences and Cell Search in LTE ................................................... 15
1.2.2.
ZadoffChu Sequences ................................................................................................ 16
1.2.3.
Primary Synchronization Signal (PSS) Sequences ...................................................... 17
1.2.4.
PSS Generation ............................................................................................................ 17
1.2.5.
Secondary Synchronization Signal (SSS) Sequences.................................................. 18
1.2.6.
Cell Search Performance .............................................................................................. 19
1.2.7.
Reference Signals and Channel Estimation ................................................................. 19
1.2.8.
Design of Reference Signals in LTE ............................................................................. 19
1.2.9.
Cell-Specific Reference Signals (CRS) ......................................................................... 20
1.2.10.
UE-Specific Reference Signals(URS) ........................................................................... 21
1.2.11.
RS-Aided Channel Modelling and Estimation ............................................................... 22
1.2.12.
Frequency Domain Channel Estimation ....................................................................... 22
1.2.13.
Time-Domain Channel Estimation ................................................................................ 22
1.2.14.
Spatial Domain Channel Estimation(SD-MMSE) .......................................................... 23
1.3
Phy Data and Control Channels - DL .................................................................................... 25
1.3.1.
Physical Broadcast Channel (PBCH) ............................................................................ 25
1.3.2.
Physical Downlink Shared Channel (PDSCH) .............................................................. 26
1.3.3.
Physical Multicast Channel (PMCH) ............................................................................. 27
1.3.4.
Downlink Control Channels ........................................................................................... 27
1.3.5.
Physical Control Format Indicator Channel (PCFICH).................................................. 28
1.3.6.
Physical Downlink Control Channel (PDCCH) .............................................................. 29
1.3.7.
PDCCH Candidate Selection ........................................................................................ 30
1.3.8.
Formats for Downlink Control Information (DCI) ........................................................... 31
1.3.9.
Physical Hybrid ARQ Indicator Channel (PHICH) ......................................................... 35
1.3.10.
Resource Allocation ...................................................................................................... 36
1.3.11.
DL Resource Allocation Rules ...................................................................................... 38
1.3.12.
Resource Allocation Bitmap examples.......................................................................... 39
1.3.13.
Uplink Grant .................................................................................................................. 40
Page 3
1.3.14.
PDCCH Transmission and Blind Decoding ................................................................... 41
1.3.15.
Enhanced PDCCH (EPDCCH) ...................................................................................... 41
EPDCCH assignment procedure .................................................................................................. 41

Mapping to resource elements ...................................................................................................... 42
Resource mapping parameters for EPDCCH ............................................................................... 43
EPDCCH formats .......................................................................................................................... 43
1.3.16.
1.4
Physical Layer Processing - DL ............................................................................................ 47
1.4.1.
Link Adaptation and Feedback Computation ................................................................ 48
1.4.2.
CQI Feedback in LTE .................................................................................................... 48
1.4.3.
Channel Coding............................................................................................................. 49
1.4.4.
Viterbi Algorithm (VA) (Example): ................................................................................. 49
1.4.5.
LTE Contention-Free Interleaver................................................................................... 51
1.4.6.
Rate-Matching ............................................................................................................... 52
1.4.7.
HARQ in LTE................................................................................................................. 53
1.4.8.
Coding for Control Channels in LTE ............................................................................. 54
1.4.9.
General structure for downlink physical channels ......................................................... 54
1.4.10.
Scrambling .................................................................................................................... 55
1.4.11.
Modulation ..................................................................................................................... 55
1.4.12.
Layer mapping............................................................................................................... 55
1.4.13.
Precoding ...................................................................................................................... 55
1.4.14.
Mapping to resource elements ...................................................................................... 56
1.5
2.
Scheduling Process - Control Channel Viewpoint ........................................................ 45
MIMO Techniques ................................................................................................................. 59
1.5.1.
Introduction to MIMO ..................................................................................................... 59
1.5.2.
Single-User (SU-) MIMO Techniques ........................................................................... 60
1.5.3.
Multi-User Techniques .................................................................................................. 63
1.5.4.
MIMO Schemes in LTE ................................................................................................. 64
1.5.5.
Single-User Schemes ................................................................................................... 64
1.5.6.
Beamforming Schemes ................................................................................................. 65
1.5.7.
What is Spatial Multiplexing? ........................................................................................ 65
1.5.8.
Precoding ...................................................................................................................... 66
1.5.9.
Cyclic Delay Diversity (CDD) ........................................................................................ 67
1.5.10.
Multi-User Schemes ...................................................................................................... 68
1.5.11.
Physical-Layer MIMO Performance .............................................................................. 70
Uplink Physical Layer .................................................................................................................... 71

2.1.
SC-FDMA Principles ............................................................................................................. 71
2.1.1.
SC-FDMA Signal Generation (DFT-S-OFDM) .............................................................. 71
2.1.2.
SC-FDMA Design parameters in LTE ........................................................................... 72
Page 4
2.2.
UL Physical Channel Structure ............................................................................................. 75
2.2.1
Uplink Shared Data Channel Structure ......................................................................... 75
2.2.2
Scheduling in LTE SC-FDMA Uplink............................................................................. 76
2.2.3
Uplink Control Channel Design ..................................................................................... 77
2.2.4
Physical Uplink Control Channel (PUCCH) .................................................................. 77
2.2.5
Multiplexing of UEs within a PUCCH Region ................................................................ 78
2.2.6
Control Signalling Information Carried on PUCCH ....................................................... 79
2.2.7
CQI Transmission on PUCCH (Format 2) ..................................................................... 79
2.2.8
Multiplexing CQI and ACK/NACK on PUCCH .............................................................. 79
2.2.9
CQI and ACK/NACK in Same RB (Mixed PUCCH RB) ................................................ 81
2.2.10
Scheduling Request (SR) on PUCCH (Format 1) ......................................................... 81
2.2.11
Control Signalling and UL-SCH multiplexing on PUSCH .............................................. 82
2.2.12
Multiple-Antenna Techniques ........................................................................................ 83
2.2.13
PUSCH UE Antenna Selection Indication ..................................................................... 83
2.2.14
Multi-User Virtual MIMO or SDMA ............................................................................... 83
2.3.
Uplink Reference Signal ........................................................................................................ 85
2.3.1.
UL RS Signal Sequence Generation ............................................................................ 85
2.3.2.
Base RS Sequences and Sequence Grouping ............................................................. 85
2.3.3.
Orthogonal RS via Cyclic Time-Shifts of a Base Sequence ......................................... 86
2.3.4.
Sequence-Group Hopping and Planning ...................................................................... 87
2.3.5.
Cyclic Shift Hopping ...................................................................................................... 88
2.3.6.
Demodulation Reference Signals (DM RS) .................................................................. 88
2.3.7.
Uplink Sounding Reference Signals (SRS) ................................................................... 89
2.4.
Uplink Capacity and Coverage.............................................................................................. 91
2.4.1
Uplink Capacity - Factors Affecting Uplink Capacity ..................................................... 91
2.4.2
Uplink Power Control and Interference Management ................................................... 92
2.4.3
Uplink Control Channel Overhead ................................................................................ 92
2.4.4
Modulation and Number of HARQ Transmissions ........................................................ 92
2.4.5
Delay Constraints and VoIP .......................................................................................... 92
2.4.6
Number of eNodeB Receive Antennas ......................................................................... 92
2.4.7
Minimum Size of Resource Allocation........................................................................... 93
2.4.8
LTE Uplink Capacity Evaluation .................................................................................... 93
2.4.9
LTE Uplink Coverage and Link Budget ......................................................................... 93
2.5.
Random Access .................................................................................................................... 95
2.6.1
Random Access Procedure .......................................................................................... 95
2.6.2
Contention-Based Random Access Procedure ............................................................ 95
2.6.3
Physical Random Access Channel Design ................................................................... 98
2.6.4
Preamble Sequence Theory and Design .................................................................... 102
Page 5
2.6.5
PRACH Implementation .............................................................................................. 108
2.6.6
Time Division Duplex (TDD) PRACH .......................................................................... 109
2.6.7
Uplink Timing Control .................................................................................................. 110
2.6.8
Timing Advance Procedure ......................................................................................... 110
2.6.9
Power Control.............................................................................................................. 111
2.6.
Miscellaneous...................................................................................................................... 116
2.6.1
Evaluation LTE Physical Layer Questions .................................................................. 116
Page 6
1. Downlink Physical Layer

1.1 OFDMA Principles
OFDMA (Orthogonal Frequency division Multiple Access) is a multicarrier scheme. Multicarrier
schemes subdivide bandwidth into parallel subchannels, ideally each non-frequency-selective
(spectrally-flat gain), overlapping but orthogonal. This avoids need of guard-bands, makeing OFDM
highly spectrally efficient, as subchannels can be perfectly separated at the receiver. This makes
receiver less complex, attractive for high-rate mobiles. Robustness has to be built up against timevariant channels by employing channel coding. LTE downlink combines OFDM with channel coding
and Hybrid Automatic Repeat reQuest (HARQ). OFDM is ideal for broadcast/DL applications for low
receiver complexity. OFDM has efficient implementation by means of the FFT. It uses Cyclic Prefix to
avoid ISI, enabling block-wise processing. Orthogonal subcarriers avoid spectrum wastage in
intersubcarrier guard-bands. Parameters flexibility allows balance the tolerance of Doppler and delay
spread.
Key OFDMA points

(a)
(b)
(c)
(d)
(e)
(f)
(g)
Orthogonal subcarriers with very small inter-subcarrier guard-bands.

It makes use of a CP to avoid ISI, enabling block-wise processing.
Efficient implementation by means of the FFT.
Achieves high transmission rates of broadband transmission, with low receiver complexity.
Balanced tolerance of Doppler and delay spread depending on the deployment scenario.
It can be extended to a multiple-access scheme, OFDMA, in a straightforward manner.
Suited for broadcast or downlink applications because of low receiver complexity while
requiring a high transmitter complexity (expensive PA).
First OFDM patent filed at Bell Labs in 1966, initially only as analog. In 1971, Discrete Fourier
Transform (DFT) was proposed. Later in 1980, application of the Winograd Fourier Transform (WFT)
or Fast Fourier Transform (FFT) was employed. OFDM then became modulation of choice for ADSL
and wireless systems.
OFDM tended to focus broadcast systems such as - Digital Video Broadcasting (DVB) and Digital
Audio Broadcasting (DAB), and WLANs. Main thing to control in OFDM was PAPR and thats why in
low power WLAN it was good. First cellular mobile based on OFDM was proposed in 1985 by IEEE
to LTE downlink. Other benefits of OFDM was to operate in different bandwidth according to spectrum
availability.
1.1.1 OFDM - Orthogonal Multiplexing Principle

Challenge is always in having a symbol period Ts < channel delay spread Td. This generates
Intersymbol Interference (ISI), needing complex equalization procedure. Equalization complexity
usually is in proportion to square of (channel impulse response length). Data symbols are first serialto-parallel converted for modulation on M parallel subcarriers.This increases symbol duration by a
factor of approx M, > channel delay spread.
Page 7
Fig 2.1.1.1 OFDM Signal Processing

This operation makes time-varying channel impulse response substantially constant during each
modulated OFDM symbol. Resulting long symbol duration is virtually unaffected by ISI compared to
the short symbol duration. A Serial to Parallel (S/P) converter collects serial data symbols into a data
block
Sk = [Sk [0] , Sk [1] , . . . , Sk [M 1]]T
of dimension M, where the subscript k is the index of an OFDM symbol (spanning the M sub-carriers).
The M parallel data streams are first independently modulated resulting in the complex vector
Xk = [Xk [0] , Xk [1] , . . . , Xk [M 1]]T .
In principle it is possible to use different modulations (e.g. QPSK or 16QAM) on each sub-carrier, the
channel gain may differ between sub-carriers, and thus some sub-carriers can carry higher data-rates
than others. The vector of data symbols Xk then passes through an Inverse FFT (IFFT) resulting in a
set of N complex time-domain samples
xk = [xk[0], . . . , xk[N 1]]T .
In a practical OFDM system, the number of processed subcarriers is greater than the number of
modulated sub-carriers (i.e. N M), with the unmodulated sub-carriers being padded with zeros.
Fig 2.1.1.2 OFDMA tramsmission and reception

A guard period is created at the beginning of each OFDM symbol, to eliminate the remaining impact
of ISI. A Cyclic Prefix (CP) is added at the beginning of each symbol xk. The CP is generated by
duplicating the last G samples of the IFFT output and appending them at the beginning of xk. This
yields the time domain OFDM symbol [xk[N G], . . . , xk[N 1], xk[0], . . . , xk[N 1]]T . CP length
G should be longer than the longest channel impulse response to be supported. The CP converts the
linear (i.e. aperiodic) convolution of the channel into a circular (i.e. periodic) one which is suitable for
DFT processing. The IFFT output is then Parallel-to-Serial (P/S) converted for transmission through
frequency-selective channel. Here is an example of OFDM LTE signal.
At the receiver, the reverse operations are performed to demodulate the OFDM signal, CP are
removed and ISI-free block of samples is passed to the DFT. If number of subcarriers N is designed
to be a power of 2, a highly efficient FFT implementation may be used to transform the signal back to
the frequency domain. Among the N parallel streams output from the FFT, the modulated subset of M
subcarriers are selected and further processed by the receiver.
Let x(t) be the signal symbol transmitted at time instant t . The received signal in a multipath
environment is then given by r(t) = x(t) * h(t) + z(t), where h(t) is the continuous-time impulse
response of the channel, represents the convolution operation and z(t) is the additive noise. Assuming
that x(t) is band-limited to [1/2Ts ,1/2Ts], the continuous-time signal x(t) can be sampled at sampling
rate Ts such that the Nyquist criterion is satisfied. Due to multipath, several replicas of the transmitted
signals arrive at the receiver at different delays. The received discrete-time OFDM symbol k including
CP, under the assumption that the channel impulse response has a length smaller than or equal to G,
Receiver has to process equalization to recover xk[n] signals. CP of OFDM changes the linear
convolution into a circular one. The circular convolution is very efficiently transformed by an FFT into
a multiplicative operation in frequency domain. Hence, the transmitted signal over a frequencyselective (multipath) channel is converted into a transmission over N parallel flat-fading channels in
Page 8
the frequency domain: Rk[m] = Xk[m] H[m] + Zk[m]. As a result the equalization is much simpler than
for single-carrier systems and consists of just one complex multiplication per subcarrier.
1.1.2 Peak-to-Average Power Ratio and Sensitivity

Major drawbacks of OFDM is that it has a high Peak-to-Average Power Ratio (PAPR). The amplitude
variations of OFDM signal can be very high, however PAs of RF transmitters are linear only within a
limited dynamic range. Hence, OFDM signal is likely to suffer from non-linear distortion caused by
clipping, giving out-of-band spurious emissions and in-band corruption of the signal. To avoid such
distortion, PAs should have large power back-offs, leading to inefficient amplification.
Let x[n] be the signal after IFFT. PAPR of an OFDM symbol is defined as the square of the peak
2
2
amplitude divided by the mean power, i.e. PAPR = Max,n{|x[n]| } / E{|x[n]| }
It is observed that a high PAPR does not occur very often. However, when it does occur, degradation
due to PA non-linearities may be expected.
PAPR Reduction Techniques
Many techniques are studied for reducing the PAPR, but not specified for downlink. An overview of
possibilities is provided below.
1. Clipping and filtering .
Signal may be clipped, but it causes spectral leakage into adjacent channels, resulting in
reduced spectral efficiency, in-band noise, degrading BER. To avoid this problem,
oversample the original signal by padding with zeros and processing it using a longer
IFFT. Oversampled signal is clipped and then filtered to reduce the out-of-band radiation.
This may be is used in LTE.
2. Selected mapping.
Whichever phase vector gives Least PAPR, that is used. To recover phase information,
separate control signalling is used to tell which phase vector was used. It is not used.
3. Coding techniques.
Use code words with lowest PAPR. Complementary codes have good properties to
combine both PAPR and forward error correction. It is not used.
Sensitivity to Carrier Frequency Offset and Time-Varying Channels

OFDM orthogonality relies that transmitter and receiver operate with exactly same frequency
reference, else perfect orthogonality of subcarriers is lost, causing subcarrier leakage (Inter-Carrier
Interference (ICI).
UE local oscillator frequency drifts are usually greater than in the eNodeB and are typically due to
temperature and voltage variation and phase noise. This difference between the reference
frequencies is referred as Carrier Frequency Offset (CFO). The CFO can be larger than subcarrier
spacing - divided into integer part and fractional part. Frequency error fo = (T+e)df. Where, df is
subcarrier spacing,, T is an integer and 0.5 <e <0.5.
If T != 0, then the modulated data are in wrong positions, resulting in BER of 0.5 if the frequency offset
is not compensated at the receiver independently of the value of e. In case of T=0 and e<>0,
perfect orthogonality is lost, resulting in ICI with BER.
Even relative speed between transmitter and receiver also generates a frequency error due to
Doppler shift fd. ICI resulting from a mismatch fo between the transmitter and receiver oscillator
frequencies can be modelled as a Doppler shift.
The sensitivity of the BER depends on the modulation order. QPSK modulation can tolerate up to e max
= 0.05(5%), whereas 64-QAM requires e 0.01(1%).
1.1.3 Timing Offset and Cyclic Prefix Dimensioning

OFDM can be insensitive to timing synchronization errors provided, the misalignment remains within
cyclic prefix duration. If To TCP (with To being the timing error), then orthogonality is maintained. Any
symbol timing delay only introduces a constant phase shift from one subcarrier to another. The
received signal at themth subcarrier is given by:
Rk[m] = Xk[m].e(j2 dm/N )
Page 9
where d is the timing offset in samples corresponding to a duration equal to To.

This phase shift can be recovered as part of the channel estimation operation, with cyclic prefix but
not zero-padding.In the general case of a channel with delay spread, for a given CP length the
maximum tolerated timing offset without degrading the OFDM reception is reduced by an amount
equal to the length of the channel impulse response: To TCP Td. For greater timing errors, ISI and
ICI occur. Timing synchronization becomes more critical in long-delay spread channels. Initial timing
is achieved by the cell-search and synchronization procedures. Thereafter, for continuous tracking of
timing-offset, either CP correlation or Reference Signals (RSs) is used.
If an OFDM system, CP is sufficiently designed of lengthG samples such that Channel impulse
Response L<G, to turn the linear convolution into a circular one to keep the subcarriers orthogonal.
1.1.4 OFDMA Parameter Dimensioning

Orthogonal Frequency Division Multiple Access (OFDMA) is multiuser OFDM, by.distributing
subcarriers to different users simultaneously. Usually, subcarriers are allocated in contiguous groups
to a user.
Based on feedback about frequency-selective channel conditions from each user, adaptive user-tosubcarrier assignment is performed, to enhance system spectral efficiency. OFDMA is combined with
TDMA, such that resources are partitioned in the time-frequency plane groups of subcarriers for a
specific duration. In LTE, such time-frequency blocks if 12 subcarriers for 1ms is one Resource Blocks
(RBs).
The main propagation characteristics (1) expected delay spread Td, (2) maximum Doppler
frequency fdmax , and, (3) targeted cell size, are considered for parameter dimensioning. This defines
the CP length and subcarrier spacing.
CP should be longer than channel impulse response for robustness against ISI. Large cells, longer
delay spreads need a longer CP. But longer CP means larger overhead (energy per transmitted bit).
Out of the N + G transmitted symbols, only N convey information, leading to a rate loss. Let TCP =
GTs and OFDM symbol period Tu = NTs , then overhead factor: overhead = TCP / (Tu + TCP).
To maximize spectral efficiency, Tu should be larger relative to CP, but small enough to ensure that
the channel does not vary within one OFDM symbol.
Further, Tu is related to subcarrier spacing by df = 1/Tu. Choosing a large Tu leads to a smaller
subcarrier separation df perfection , which impact sensitivity to Doppler and frequency offset.
Thus, in summary, three design criteria are identified:
TCP Td
Fdmax/df << 1
TCP * df << 1
to prevent ISI,
to keep ICI due to Doppler sufficiently low,
for spectral efficiency.
1.1.5 Physical Layer Parameters for LTE

LTE needs to support speed 350 or even 500 km/h, cell sizes may range from few metres to large
cells with radii of tens of kilometres, carrier frequencies from 400MHz to 4 GHz, bandwidths ranging
from 1.4 to 20 MHz with different delay spreads and Doppler frequencies.
In general, LTE has devised following 3 types of CP duration.
1. The normal LTE downlink uses a df = 15 kHz subcarrier spacing with CP length of
approximately 5 s. This is a compromise between % overhead of CP and sensitivity to
frequency offsets. A 15 kHz spacing is sufficiently large to allow high mobility and avoid need
for closed-loop freq adjustments.
2. In addition, it is possible to configure with an extended CP of approximately 17 s, designed
for large cells to ensure the delay spread should be contained within CP, of course with
higher overhead.
3. LTE is also designed to support Multimedia Broadcast Single Frequency Network (MBSFN),
where UE receives and combines synchronized signals from multiple cells. In this case,
relative timing offsets from multiple cells must all be received within CP duration. To avoid a
Page 10
further overhead, sub-carrier spacing is kept at 7.5kHz and an extended CP of approx 33 s

is used.
Fig 2.1.5.1 -FDD Frame Structure
Fig 2.1.5.2 TDD Frame Structure

Note that, with normal CP, the CP for the first symbol in each 0.5 ms slot is slightly longer than the
next six symbols, to accommodate an integer (7) number of symbols in each slot, with assumed FFT
block-lengths of 2048. For 20 MHz, FFT order of 2048 is assumed for efficient implementation.
However, in practice the implementer is free to use other Discrete Fourier Transform sizes.
These parameterizations are designed to be compatible with a sampling frequency of 30.72 MHz,
which is 8*3.84Mhz(UMTS sampling rate), for backward compatibility. Thus, the basic unit of time in
LTE, is defined as Ts = 1/30.72 s. Lower sampling frequencies (and proportionally lower FFT orders)
are always possible to reduce RF and baseband processing complexity for narrower BW: Example,
for 5 MHz, FFT order and sampling frequency could be 512 and fs = 7.68 MHz respectively, while
only 300 subcarriers are actually modulated with data.
For simple implementation, direct current (d.c.) subcarrier is left unused, to avoid d.c. offset errors.
1.1.6 Transmission Resource Structure

LTE downlink, consist of user-plane and control-plane data from higher protocol stack layers
multiplexed with physical layer signalling. A DL resources possess dimensions of time(slot),
Page 11
frequency(multiple of 180khz) and space(layer). Layer is defined by multiple antenna transmission

and reception.
The largest unit of time is the 10 ms radio frame, further subdivided into ten 1 ms subframes, each of
which is split into two 0.5 ms slots. Each slot has seven OFDM symbols in normal CP (six if extended
CP). In frequency domain, resources are grouped in units of 12 subcarriers (15*12kHz=180 kHz),
such that one unit of 12 subcarriers for a duration of one slot is termed a Resource Block (RB).
Fig 2.1.6.1 FDD Downlink Frame sample

Smallest unit of resource is the Resource Element (RE) - one subcarrier for a duration of one OFDM
symbol. A RB comprised of 84 REs in normal CP (72 RE in extended CP). Within certain RBs, some
REs are reserved for synchronization signals (PSS/SSS), reference signals (RS), control signalling
and critical broadcast system information (CFICH,PHICH,PDCCH). Remaining REs are used for data
transmission(PDSCH), and are usually allocated in pairs (in time domain) of RBs.
Fig 2.1.6.2 TDD Downlink Frame sample

Two types of frame structure are defined:
1. Frame Structure Type 1(Frequency Division Duplex,FDD) assumes all subframes are
available for DL, in paired radio spectrum, or standalone downlink carrier.
2. Frame Structure Type 2(Time Division Duplexing , TDD) in unpaired spectrum, basic
structure of RBs and REs remains same, but only a subset of subframes are available for
Page 12
downlink; remaining subframes are used for uplink or for special subframes which allow
switching between DL & UL. In the centre of the special subframes a guard period is provided
which allows UL timing to be advanced.
Signal Structure
Physical layer translate data into reliable signal for transmission between eNodeB and UE. Each
block of data is first protected against transmission errors, first with a Cyclic Redundancy Check
(CRC), and then with coding; The initial scrambling stage is applied to all DL channels and helps
interference rejection. Scrambling sequence uses order- 31 Gold code, which are not cyclic shifts of
each other.
Scrambling sequence generator is re-initialized every subframe (except PBCH), based on cell-id,
subframe number (within a radio frame), UE identity and codeword id.
Scrambling sequence generator is similar to pseudo-random sequence used for Reference Signals,
only difference is the method of initialization. A fast-forward of 1600 places is applied at initialization to
ensure low cross-correlation between sequences used in adjacent cells.
Following scrambling, data bits from each channel are mapped to modulation symbols depending on
modulation scheme, then mapped to layers, precoded, mapped to RE, and finally translated into a
complex-valued OFDM signal by IFFT.
To communicate with eNodeB cells, UE must first identify the DL from one of these cells and
synchronize with it. This is achieved by means of special synchronization signals embedded into the
OFDM structure by cell search and synchronization. Then UE estimates DL radio channel to perform
demodulation of received DL signal, based on pilot signals (reference signals) inserted into DL signal.
The channel designs are explained next.
Page 13
Page 14
1.2 Synchronization and Cell Search

Cell Search executes synchronization for time and frequency parameters, necessary to demodulate
DL and to transmit UL with correct timing and acquires some critical system parameters.
Three major synchronization requirements:
1. symbol timing determines correct symbol start position and sets the FFT window position;
2. carrier frequency synchronization to reduce frequency errors by oscillator & Doppler shift;
3. sampling clock synchronization.
1.2.1. Synchronization Sequences and Cell Search in LTE

Two relevant cell search procedures exist in LTE:
1. Initial synchronization,
UE detects a cell and decodes all information required to register. This is required, for
example, when UE is switched on, or it has lost connection to the serving cell.
2. New cell identification,
When UE is already connected and is detecting a new neighbour cell. UE reports new
cell measurements to Serving cell for handover. Procedure is repeated periodically
until either Scell quality becomes satisfactory again, or UE moves to another cell.
Fig 2.2.1.1 FDD and TDD Synchronization Signalling

Synchronization procedure detects specially designed Primary Synchronization Signal (PSS) and
Secondary Synchronization Signal (SSS). This enables time and frequency synchronization, provides
the UE with physical cell identity (PCI) and CP, and informs UE whether cell uses FDD or TDD. Here
is figure explaining relative location of PSS and SSS in frame structure of FDD and TDD respectively.
Fig 2.2.1.2 Cell Synchronization Process

In initial synchronization, UE proceeds to decode PBCH for critical system information (SI). For new
cell identification, UE does not need to decode PBCH; it makes quality-level measurements
(RSRP/RSRQ) and reports to the serving cell.
Page 15
The sync signals are transmitted periodically, twice per 10 ms radio frame.
th
th
In FDD cell, PSS is always located in the last symbol of the 0 and 10 slots of each frame,
thus enabling UE to acquire the slot boundary timing independently of the CP. SSS is located
in the symbol immediately preceding PSS, for coherent detection of SSS relative to PSS.
nd
nd
th
In TDD cell, PSS is located in 2 symbol of the 2 and 12 slots, while the SSS is located 3
symbols earlier and falls in previous slot.
Position of SSS changes depending on CP length of the cell. At this stage, CP length is unknown and
SSS is blindly detected by checking for SSS at expected positions.
PSS in a given cell is same in every subframe, SSS may change & thus UE knows the position of the
10 ms radio frame boundary. PSS and SSS are transmitted in the central six Resource Blocks (RBs),
irrespective of the system BW (6 to 110 RBs), without knowing BW. The PSS and SSS are each
comprised of a sequence of length 62 symbols, mapped to the central 62 subcarriers around d.c.
subcarrier which is left unused. Five REs at each extremity of each sync sequence are not used. Thus
a UE can detect the PSS and SSS with size-64 FFT and a lower sampling rate if all 72 subcarriers
were used.
In case of MIMO at eNodeB, PSS and SSS are always transmitted from same antenna port in a
subframe, while between different subframes they may be transmitted from different antenna ports for
diversity.
PSS and SSS sequence indicate one of 504 unique PCI, grouped into 168 groups of three identities.
The three identities in a group are assigned to cells under same eNodeB. Three PSS sequences are
used to indicate the cell identity within the group, and 168 SSS sequences are used to indicate the
identity of the group.
PSS uses ZadoffChu sequences
1.2.2. ZadoffChu Sequences

ZadoffChu (ZC) sequences (Generalized Chirp-Like (GCL) sequences) are non-binary unitamplitude sequences, which satisfy a Constant Amplitude Zero Autocorrelation (CAZAC) property.
The ZC sequence of odd-length NZC is given by
aq(n) = exp [j2q (n(n + 1)/2 + ln)/ NZC ]
where q {1, . . . , NZC 1} is the ZC sequence root index, n = 0, 1, . . . , N ZC 1, l N is any integer
(In LTE l = 0).
ZC sequences have the following important properties.
Property 1. A ZC sequence has constant amplitude which limits PAPR and generates bounded and
time-flat interference to other users, and its NZC-point DFT also has constant amplitude.
Property 2. ZC sequences of any length have ideal cyclic autocorrelation (correlation with circularly
shifted version of itself is a delta function). ZC periodic autocorrelation is exactly zero for <> 0 and it
is non-zero for = 0, whereas PN periodic autocorrelation shows significant peaks, some above 0.1,
at non-zero lags.
CAZAC sequence allows multiple orthogonal sequences to be generated from the same ZC
sequence. Indeed, if the periodic autocorrelation of a ZC sequence provides a single peak at the zero
lag, the periodic correlation of the same sequence against its cyclic shifted replica provides a peak at
lag NCS, where NCS is the number of samples of the cyclic shift. This creates a Zero-Correlation Zone
(ZCZ) between the two sequences. As a result, as long as the ZCZ is dimensioned to cope with the
largest possible expected time misalignment between them, the two sequences are orthogonal for all
transmissions within this time misalignment.
Property 3. The absolute value of the cyclic cross-correlation function between any two ZC
sequences is constant and equal to 1/NZC, if |q1 q2| (where q1 and q2 are the sequence indices) is
relatively prime with respect to NZC . Selecting NZC as a prime number results in NZC 1 Zaddoff-Chu
sequences which have the optimal cyclic cross-correlation between any pair. Cyclic extension or
truncation preserves both the constant amplitude property and the zero cyclic autocorrelation property
for different cyclic shifts.
The DFT of a ZC sequence xu(n) is a weighted cyclicly-shifted ZC sequence Xw(k) such that w = 1/u
mod NZC. This means that a ZC sequence can be generated directly in the frequency domain without
the need for a DFT operation.
Page 16
1.2.3. Primary Synchronization Signal (PSS) Sequences

There are 0 to 503 (total 504) PCI are available to be assigned. Every cell will have a PCI where
PCI=3* NID,1 + NID,2.
The NID,1 is the PCI group ID - , NID,1 can have values 0 to 167 and
NID,2 is the PCI local (may be sector) ID - NID,2 can have values 0 to 2.
How does the PCI optimization affect my Network? Well, this is the main parameter, by which the
PSS, SSS and reference signals will be generated. Even your scrambling code with which every DL
and UL signal will be scrambled, will depend on this. So, every generated signals uniqueness
depends on this parameter. Lets understand how some of the signals are generated based on PCI.
If PSS, SSS, RS and other generated signals are not unique, then my every operation will be affected
and it may reflect as latency in Synchronization detection, Interference and lower SINR values for
signals, which will end up in low CQI.
1.2.4. PSS Generation

The PSS represented by d(u,n) is generated from a frequency-domain Zadoff-Chu sequence
according to
j un( n 1)
63
e
d u (n) u ( n 1)( n 2)
e j
63
n 0,1,...,30
n 31,32,...,61
where the Zadoff-Chu root sequence index u is given by following table.
Root index u
(2)
N ID
0
1
2
25
29
34
The mapping of PSS to resource elements depends on the frame structure, FDD or TDD. The
sequence d(u,n) is mapped to the resource elements according to
ak ,l d n ,
k n 31
n 0,...,61
DL RB
N RB
N sc
Page 17
For FDD, the PSS is mapped to the last OFDM symbol in slots 0 and 10. For TDD, the PSS is apped
to the third OFDM symbol in subframes 1 and 6. Resource elements (k,l) in the OFDM symbols used
for transmission of the primary synchronization signal where
DL RB
N RB
N sc
2
n 5,4,...,1,62,63,...66 are reserved and not used for transmission of the primary synchronization
signal. N=0,1,61 are used with above formulae.
k n 31
PSS is constructed from a freq-domain ZC sequence of length 63, with middle element punctured to
avoid transmitting on d.c. subcarrier.
This set of roots for ZC sequences was chosen for its good periodic autocorrelation and crosscorrelation properties. These sequences have a low-frequency offset sensitivity (maximum undesired
autocorrelation peak /desired correlation peak) at a certain frequency offset, giving best robustness.
Also the ZC sequences are robust against frequency drifts. Thus, PSS can be easily detected during
the initial synchronization with a frequency offset up to 7.5 kHz.
The selected root combination satisfies time-domain root-symmetry, sequences 29 and 34 are
complex conjugates of each other and can be detected with a single correlator. UE must detect PSS
without any prior knowledge of the channel, so noncoherent correlation is required for PSS timing
detection.
1.2.5. Secondary Synchronization Signal (SSS) Sequences

SSS maximum length M-sequences, can be created by cycling through every possible state of a shift
register of length n, resulting in M-sequence of length 2n 1.
Two length-31 BPSK secondary sync codes (SSC1(even (di)) and SSC2(odd d(i)) are interleaved to
construct SSS sequence in frequency-domain. Two codes are two different cyclic shifts of a single
length-31 M-sequence. Cyclic shift indices are derived from a function of PCI group. Two codes are
alternated between the first and second SSS in each radio frame.
(m )
m0 m mod 31
s 0 (n)c0 n in subframe 0
d (2n) 0( m )
1
s1 (n)c0 n in subframe 5
with m1 m0 m 31 1 mod 31
s1( m1 ) (n)c1 n z1( m0 ) n in subframe 0
N (1) q(q 1) 2
d (2n 1) ( m )
(1)
(m )
m N ID(1) q(q 1) 2 , q ID
, q N ID 30
s0 0 (n)c1 n z1 1 n in subframe 5
30
Thus UE determines the 10 ms radio frame timing from a single observation of a SSS. SSC2 is
scrambled by a sequence that depends on the index of SSC1. Sequence is then scrambled by a code
that depends on the PSS. Scrambling code is mapped to the PCI within the group corresponding to
the target eNodeB.
The resource mapping is done as per the following:
Page 18
ak ,l d n ,
k n 31
n 0,...,61
DL RB
N RB
N sc
2
N DL 2 in slots 0 and 10 for frame structure type1

l symb
DL
N symb 1 in slots1 and 11 for frame structure type 2
SSS sequences are spectrally flat. PSS, the SSS can be detected with a frequency offset up to 7.5
kHz. Channel is known based on the PSS sequence first and then SSS detection is done.
However, in the case of synchronized neighbouring eNodeBs, coherent detector performance can be
degraded. If an interfering eNodeB employs the same PSS, phase difference between them can have
adverse impact on estimation of the channel coefficients. If BW of the channel is less than the six RB
for SSS, impact may be bad, hence minimum 6 RB legth is chisen. M-sequence and Walsh
Hadamard matrices are similar and index remapping is done. This reduces complexity of SSS
detector, as complexity = N log2 N with N=32, complexity= 32 log2 32 = 160.
1.2.6. Cell Search Performance

A new cell detection delay for UE and report it to S-eNB should be less than acceptable threshold.
eNodeB uses the reports to prepare intra- or inter-frequency handover.
A multicell environment with three cells with different transmitted powers with synchronized and
unsynchronized eNodeBs should be analysed. For the propagation channel, various multipath fading,
at least two receive antenna, UE speeds, (5 km/h, 300 km/h) should be considered.
Cell search performance is measured as 90-percentile(maximum time required to detect a target cell
90%of the time) identification delay. After detection of PSS-SSS, RSRP is measured. For initial
synchronization case the time taken to decode PBCH is adapted, and not just of reporting of
measurements on RS.For inter-frequency handover, performance can be derived from the intrafrequency performance timing.
Coherent Versus Non-Coherent Detection

A coherent detector uses knowledge of the channel, while a non-coherent detector uses an
optimization metric of average channel statistics. In PSS, non-coherent (No channel estimation
available) detection is used, while for SSS, coherent (channel estimation) or non-coherent techniques
can be used.
1.2.7. Reference Signals and Channel Estimation

In any communication system signal x transmitted by A passes through a radio channel H (exhibit
multipath fading, causing ISI) and suffers additive noise before being received by B. To remove ISI,
equalization, detection algorithms and knowledge of Channel Impulse Response (CIR) is used.
OFDMA is quite robust against ISI by CP which allows very good equalization at receiver.
Coherent detection uses amplitude and phase information exchanged between eNodeBs and UEs.
This comes at a price of overhead of channel estimation by exploiting known signals which do not
carry any data, sacrificing spectral efficiency. Known reference signals are inserted into the
transmitted signal structure. Reference signals(known) are multiplexed with data symbols (unknown
at receiver) in either frequency, time or code domains. Time multiplexing, known preamble-based
training transmission also is another technique.
Orthogonal RS multiplexing is the most common technique. OFDM transmission is a two-dimensional
lattice in time and frequency, which helps multiplexing of RSs mapped to specific REs according to a
specific pattern. Since RS are sent only on particular OFDM REs (particular symbols and subcarriers),
channel estimates for non-RS REs have to be computed via interpolation.
1.2.8. Design of Reference Signals in LTE

In DL, three different types of RS are provided:
Page 19
1. Cell-specific RSs ( common RSs).

2. UE-specific RSs, in the data for specific UEs.
3. MBSFN-specific RSs, for MBSFN.
1.2.9. Cell-Specific Reference Signals (CRS)

In an OFDM-based system, an equidistant arrangement of RS in the lattice structure diamond shape
achieves the Minimum Mean Squared Error (MMSE) estimate.
In time domain, RS spacing in is governed by maximum Doppler spread (highest speed540km/h(150m/s)) to be supported. Doppler shift is fd = (fc v/c) where fc is the carrier frequency, v is
8
9
UE speed in m/s, and c= 3 * 10 m/s. Considering fc = 2 GHz (2*10 Hz) and v = 500 km/h, fd
9
8
=(2*10 *150/3*10 ) =1000 Hz. According to Nyquists sampling theorem, minimum sampling
frequency to reconstruct the channel is Tc = 1/(2fd)= 0.5 ms. This implies that two RS/slot are needed
to estimate channel correctly.
Fig 2.2.8.1 Cell RS for 1, 2 and 4 antenna

In frequency domain, there is one RS every six subcarriers on each symbols including RS symbol, but
staggered so that within each RB there is one RS every 3 subcarriers. This spacing is goverened by
expected coherence BW of channel, governed by channel delay spread. The 90% and 50%
coherence BW are given respectively by Bc,90% = 1/50d=20kHz and Bc,50% = 1/5d=200kHz where
d is the r.m.s delay spread=1000ns. In LTE spacing between two RS in frequency, is 45 kHz (3
symbols), enough to resolve expected frequency domain variations of the channel.
RS patterns are designed to work with MIMO antennas defined for multiple antenna ports at
eNodeB. An antenna port may be either a single physical antenna, or a combination of multiple
physical antenna elements. The transmitted RS in a given antenna port defines the antenna port from
the point of view of the UE, and enables UE to derive channel estimate for that antenna port.
Page 20
Fig 2.2.8.2 Antenna ports and Physical antennas

Up to eight cell-specific antenna ports may be used by eNodeB, requiring UE to derive up to eight
separate channel estimates. For each antenna port, a different RS pattern is designed, to minimize
intra-cell interference between multiple transmit antenna ports. Rp is used for RS Tx on antenna port
p. Also, when a RE is used for RS on one antenna port, corresponding RE on other antenna ports is
set to zero to limit interference. Mark that, density of RS for third and fourth antenna ports is half of
the first two, to reduce overhead. In cells with a high prevalence of high-speed users, use of four
antenna ports is unlikely, RSs with lower density can provide sufficient channel estimation accuracy.
Fig 2.2.8.3 Antenna port example of port 0 and port 5

All the RSs (cell-specific, UE-specific or MBSFN specific) are QPSK modulated to ensure low PAPR.
The signal can be written as r(l,ns,m) = 1/2[1-2c(2m)] + j1/2[1-2c(2m+1)] where m is RS index, ns
= slot number and l =symbol number within slot, c(i) is length-31 Gold sequence, with different
initialization values depending on type of RSs. RS sequence carries unambiguously one of the 504
different cell identities, Ncell ID. For the cell-specific RSs, a cell-specific frequency shift (Ncell ID mod 6)
is also applied. This shift avoids collisions between common RS from up to six adjacent cells.
Transmission power of RS is boosted, up to max 6 dB relative to surrounding data symbols, designed
to improve channel estimation. If adjacent cells also transmit high-power RS on same REs,
interference will prevent the gain.
1.2.10. UE-Specific Reference Signals(URS)

UE-specific RS may be used in addition to CRSs, embedded only in a specific UEs scheduled RBs,
using a distinct antenna port. UE is expected to use them to derive the channel estimate for
demodulating data in PDSCH RBs.
Page 21
Fig 2.2.9 Port 5, UE specific Reference Signals

A typical usage of URSs is beam-forming of data transmissions to specific UEs. Rather than using
physical antennas of other CRS antenna ports, eNodeB may use a correlated array of antenna
elements to generate a narrow beam in the direction of a particular UE. Beam will experience different
channel response requiring URSs to demodulate the beamformed data coherently.
The URS structure is chosen not to collide with CRSs, and hence URSs does not affect CRSs. URSs
have a similar pattern as CRSs allowing re-use similar channel estimation algorithms. Density is half
of CRS, minimizing the overhead.
1.2.11. RS-Aided Channel Modelling and Estimation

Channel estimation problem is related to physical propagation, number of transmit and receive
antennas, BW, frequency, cell configuration and relative speed.
1. frequencies and BW determine the scattering.
2. Cell deployment governs multipath, delay spread and spatial correlation.
3. Relative speed sets time-variations.
Propagation conditions characterize the channel in three dimension (frequency, time and spatial)
domains. Each MIMO multipath channel component can experience different scattering conditions
across the three domains. LTE specifications do not mandate any specific channel estimation
technique, and there is therefore complete freedom in implementation provided that the performance
requirements are met and the complexity is affordable.
1.2.12. Frequency Domain Channel Estimation

The natural approach to estimate the whole CTF is to interpolate its estimate between the reference
symbol positions. As a second straightforward approach, the CTF estimate over all subcarriers can be
obtained by IFFT interpolation.
More elaborate linear estimators derived from both deterministic and statistical viewpoints are
proposed -Least Squares (LS), Regularized LS, Minimum Mean-Squared Error (MMSE) and
Mismatched MMSE. It is seen that IFFT and linear interpolation methods yield lowest performance.
The regularized LS and the mismatched MMSE perform exactly equally. Optimal MMSE estimator
outperforms any other estimator. MMSE-based channel estimation suffers the least band-edge
degradation, while all the other methods presented are highly adversely affected.
1.2.13. Time-Domain Channel Estimation

Time-Domain (TD) channel estimation requires sufficient memory for buffering soft values of data over
several symbols while the channel estimation is carried out. However, correlation between
consecutive symbols decreases as UE speed increases. TD correlation is inversely proportional to the
UE speed sets a limit on the possibilities for TD filtering in high-mobility conditions.
Finite and Infinite Length MMSE-(TD-MMSE)

The statistical TD filter which is optimal in terms of Mean Squared Error (MSE) can be approximated
in the form of a finite impulse response filter. It can be observed that, unlike Frequency-Domain (FD)
MMSE filtering, the size of the matrix to be inverted for a finite-length TD-MMSE estimator is
independent of the channel length L but dependent on the chosen FIR order M. Similarly to the FD
counterpart, the TD-MMSE estimator requires knowledge of the PDP, the UE speed and the noise
variance.
Page 22
Normalized Least-Mean-Square(NLMS)
An adaptive estimation approach can be considered which does not require knowledge of secondorder statistics of both channel and noise. A feasible solution is the Normalized Least-Mean-Square
(NLMS) estimator. It can be observed that the TD-NLMS estimator requires much lower complexity
compared to TD-MMSE as no matrix inversion is required, as well as not requiring any a priori
statistical knowledge.
Other adaptative approaches could also be considered such as Recursive Least Squares (RLS) and
Kalman-based filtering.
1.2.14. Spatial Domain Channel Estimation(SD-MMSE)

LTE UE is designed for MIMO. Consequently, whenever the channel is correlated in the spatial
domain, the correlation can be exploited to provide a further means for enhancing the channel
estimate. If it is desired to exploit spatial correlation, a natural approach is again offered by Spatial
Domain (SD) MMSE filtering.
Page 23
Page 24
1.3 Phy Data and Control Channels - DL

1.3.1. Physical Broadcast Channel (PBCH)
Basic system information (SI) carries configuration info via Broadcast Channel (BCH). The SI is
divided into two categories:
1. Master Information Block (MIB), carries limited most frequently transmitted parameters
essential for initial access to the cell, is carried on Physical Broadcast Channel (PBCH).
2. System Information Blocks (SIBs) multiplexed together with Physical Downlink Shared
Channel(PDSCH).
PBCH requires to be:
1. Detectable without prior knowledge of system bandwidth;
2. Low system overhead;
3. Reliable reception right to the edge of the LTE cells;
4. Decodable with low latency and low impact on UE battery life.
Fig 2.3.1 PBCH transmission

Detectability without prior knowledge of BW is achieved by mapping PBCH only on central 72
subcarriers of OFDM signal, regardless of actual BW. UE will have first identified the centre-frequency
synchronization signals.
Low system overhead is achieved by deliberately keeping information on PBCH to a minimum (MIB
is just 14 bits), and, since it is repeated every 40 ms, it is just 0.35 kbps.
Page 25
Reliability of PBCH is achieved with time diversity, FEC coding and antenna diversity. Time diversity
is exploited by spreading transmission of each MIB on PBCH over 40 ms ensuring reception despite
loosing one transmission. The FEC coding for the PBCH uses a convolutional coder, as bits to be
coded is small. Basic code rate is 1/3, after which a high degree of repetition of systematic bits and
parity bits is used, such that each MIB is coded at a very low code-rate (1/48 over a 40 ms period) to
give strong error protection.
PBCH uses dual-antenna receive diversity enabling wider cell coverage with fewer cell sites. Transmit
antenna diversity may be also employed at eNodeB to further improve coverage.
The exact REs used by PBCH is independent of transmit antenna ports; REs used for RS are avoided
by PBCH. Number of transmit antenna ports used by eNodeB must be determined blindly by the UE.
Discovery of number of transmit antenna ports is helped by CRC on each MIB which is masked with a
codeword representing the number of transmit antenna ports.
Low latency and a low impact on UE battery life is also facilitated by low code rate with repetition.
Full set of coded bits are divided into four subsets, each is self-decodable, which are sent in one of
four different frames during the 40 ms. UE may decode the MIB correctly from the transmission in less
than four radio frames, then UE does not need to receive other parts of PBCH in the remainder of 40
ms. On the other hand, if SIR is low, UE can receive further parts of MIB, soft-combining each part,
until successful decoding is achieved.Timing of 40 ms interval is not indicated explicitly to UE; it is
determined by scrambling and bit positions. UE can initially do four separate decodings of the PBCH
and checking the CRC for each decoding to determine 40ms boundary.
A simple approach is to perform decoding using soft combination of the PBCH over four radio frames,
advancing 40 ms sliding window one radio frame at a time until the window aligns with 40 ms period
of the PBCH and the decoding succeeds.
1.3.2. Physical Downlink Shared Channel (PDSCH)

PDSCH is the main DL data channel. It is used for all user data, broadcast SI apart from PBCH, and
paging messages. Data is transmitted on PDSCH in units of transport blocks(TB), each of which is a
MAC-PDU. TB is passed from MAC to physical layer once per Transmission Time Interval (TTI),
where a TTI is 1 ms, a subframe duration.
General Use of the PDSCH
one or, at most, two TB can be transmitted per UE per subframe, depending on transmission mode
selected for the PDSCH for each UE:
Transmission Mode 1: Single eNodeB antenna port;
Transmission Mode 2: Transmit diversity;
Transmission Mode 3: Open-loop spatial multiplexing;
Transmission Mode 4: Closed-loop spatial multiplexing;
Transmission Mode 5: Multi-user Multiple-Input Multiple-Output (MU-MIMO);
Transmission Mode 6: Closed-loop rank-1 precoding;
Transmission Mode 7: Transmission using UE-Specific RS.
Transmission Mode 8: Transmission spatial multiplexing 2 layers
Transmission Mode 9: Transmission using 8 antenna, upto 8 layers
Transmission Mode 10: Transmission using 8 antenna, upto 8 layers
Except Tx-mode 7, RS for demodulating PDSCH is given by CRS. Number of eNodeB antenna ports
for PDSCH is same as PBCH. In Tx-mode 7, UE-specific RSs provide phase reference for the
PDSCH. Tx-mode also affects DL control signalling, and CQI from UE.
After coding and mapping to spatial layers, coded data bits are mapped to modulation symbols
depending on radio channel conditions and data rate required. Modulation order may be between two
bits per symbol (QPSK) and six bits per symbol (64QAM).
The RE for PDSCH can be any which are not reserved for other purposes (i.e. RS, PSS, SSS, PBCH
and control signalling(PCFICH,PHICH,PDCCH)). When UE is given a pair of RB of PDSCH, in a
subframe, only the available RE within RB can carry PDSCH data. Allocation of RB to PDSCH for a
Page 26
UE is signalled by dynamic control signalling at the start of the relevant subframe using Physical
Downlink Control Channel (PDCCH). The mapping of data to RB is carried out in one of two ways:
1. localized mapping and
2. distributed mapping.
Localized resource mapping allocates all REs in a pair of RB to same UE.
Distributed resource mapping separates two PRB in frequency giving frequency diversity for small
amounts of data. Up to two pairs of RB may be transmitted to a UE in this way.
Example of distributed Mapping: In Voice-over-IP (VoIP) service, certain frequency resources may
be persistently-scheduled, on a periodic basis to a specific UE by RRC signalling rather than
PDCCH. As data per UE for VoIP is small (one or two pairs of RB), degree of frequency diversity
obtainable via localized scheduling is very limited. When dynamic channel-dependent PDCCH
scheduling is not done, frequency diversity is achieved through distributed mapping. A frequency-hop
occurs at slot boundary in the middle of subframe, block of UE data transmitted on one RB in first half
of subframe and on a different RB in the second half.
The potential number of VoIP users which can be accommodated in a cell increases by distributed
mapping, compared to localised.
Special Uses of the PDSCH
Apart from normal user data transmission, PDSCH is used for Dynamic BCH, all SIBs, not carried on
PBCH. The RBs for SIBs are indicated by PDCCH, same way as for other PDSCH data without any
specific UE identity, but is, rather by fixed SI-RNTI(FFFF), known to all UEs..
Another special use is for Paging, as no separate physical channel provided. Normal PDCCH
signalling is used to carry equivalent of a WCDMA paging indicator, with detailed paging information
carried on PDSCH in a RB indicated by PDCCH, using single fixed identifier, P-RNTI (FFFE).
Different UEs monitor different subframes for their paging messages with their Paging Frame and
Paging Occasion calculations.
1.3.3. Physical Multicast Channel (PMCH)

All UEs must be aware of possible existence of MBMS at physical layer. Basic structure of PMCH is
very similar to PDSCH. PMCH is designed for single-frequency network operation, whereby multiple
cells transmit the same modulated symbols with very tight time-synchronization, so that signals from
different cells are received within CP, called MBSFN (MBMS Single Frequency Network) operation.
The key differences between PDSCH & PMCH are as follows:
a. PDCCH and PHICH cannot occupy more than two OFDM symbols in MBSFN
subframe. PDCCH is used only for UL resource grants and not for the PMCH, as
scheduling of MBSFN data is carried out by higher-layer signalling.
b. RS symbols embedded in PMCH is different from PDSCH.
c. The extended CP is always used. If non-MBSFN subframes use the normal CP,
normal CP is also used in OFDM symbols used for control signalling at the start of
each MBSFN subframe.
Some spare time samples usage is unspecified between the end of the last control signalling symbol
and the first PMCH symbol, PMCH remaining aligned with the end of the subframe; eNodeB may
transmit an undefined signal or alternatively switch off its transmitter UE cannot assume anything
about transmitted signal during these samples.
A UE measuring a neighbouring cell does not need to know the allocation of MBSFN subframes,
since UE knows that the first two OFDM symbols in all subframes use the same CP and RS pattern.
The MBSFN subframes patterns in a cell is indicated in SI, which indicates if pattern of MBSFN
subframes in neighbouring cells is same or different from current cell. If different pattern, then UE can
only ascertain the pattern by reading SI of that cell.
1.3.4. Downlink Control Channels

Requirements for Control Channel Design
Control channels convey physical layer signals or messages which cannot be carried efficiently or
quickly by higher layers. The UL resources on PUSCH is determined dynamically by UL scheduling in
eNodeB, and therefore signalling must be provided to indicate to UEs which resources are granted
permission to use, together with modulation and code rate. To facilitate efficient operation of HARQ
Page 27
and ensure appropriate power levels, further physical layer signals are needed to convey ACK/NACK
by eNodeB, and power control commands.
Flexibility, Overhead and Complexity
LTE allows operation in BW from six RB (1.08MHz) to 110 RB (19.8MHz). It is also designed to
support very few users with high data rates, or very many users with low data rates. Both UL grants
and DL allocations could be required for every UE in each subframe, and may be only one RB each,
worst case.
Control channel and HARQ is designed to minimize unnecessary overhead and power saving,
scaleability and flexibility without undue decoding complexity.
Coverage and Robustness
If control channels reception fails, corresponding data transmission will also fail, and will impact
throughput efficiency. Channel coding and frequency diversity can be used to make control channels
robust. It is important to adapt transmission parameters of the control signalling for different UEs or
groups of UEs, so that lower code rates and higher power levels are only applied for only relevant
UEs as necessary (e.g. near cell border).
System-Related Design Aspects

LTE control channel is designed in a cell for a particular UE (or in some cases a group of UEs). To
minimize latency, control channel transmission should be completed within one subframe, it must be
possible to transmit multiple control channels within a subframe. Both common and dedicated control
channel messages are supported. If the data arrives at the eNodeB on a regular basis, as VoIP, it can
be controlled by persistent scheduling.
Control Channel Structure and Contents

Downlink control channels (PDCCH) can be configured to occupy the first 1, 2 or 3 OFDM symbols in
every subframe, over entire BW. There are two special cases: in subframes containing MBSFN, there
may be 0, 1 or 2 symbols, while for narrow BW (BW< 10 RB), PDCCH may be 2, 3 or 4 to ensure
sufficient coverage at cell border. This can adjust overheads of particular configuration, traffic
scenario and channel conditions.
1.3.5. Physical Control Format Indicator Channel (PCFICH)

PCFICH carries CFI indicating the number symbols used for control in each subframe. UE could
deduced CFI blind decoding each possible number of symbols, but at a cost of significant processing
load. For MBSFN carriers, there may not be any physical control channels, so PCFICH is not present.
Fig 2.3.5.1 PCFICH REG requirements

CFI values od 1,2,3 are used and 4 is reserved for future. To make it robust, each CFI is coded with
32 bit long codeword, mapped to 16 REs with QPSK modulation.
Page 28
Fig 2.3.5.2 PCFICH Mapping on neighbour Cells

PCFICH is transmitted on same ports as PBCH, with transmit diversity if more than one antenna port
is used. For frequency diversity, 16 REs are distributed across BW with a predefined pattern in the
first symbol in each DL subframe, so that UEs can always locate. This is prerequisite to decode rest
of the control signalling.
A cell-specific (PCI based) frequency offset is applied to the positions of PCFICH REs. In addition, a
cell-specific scrambling sequence (PCI based) is applied to the CFI codewords to uniquely be sent.
1.3.6. Physical Downlink Control Channel (PDCCH)

PDCCH carries many Downlink Control Information (DCI) messages in each subframe, for resource
assignments and other controls for a UE or group of UEs. Each PDCCH is transmitted using one or
more Control Channel Elements (1CCE=9REG), each CCE corresponds to nine sets of four RE
known as Resource Element Groups (1REG=4RE) . Four QPSK symbols are mapped to each REG.
The RS RE are not included within REGs. REGs concept is used for PCFICH and PHICH as well.
Fig 2.3.6.1 REG Counting

Four PDCCH formats 0,1,2,3 are supported, based on 1,2,4 or 8 CCEs are used respectively for each
PDCCH.
Page 29
Fig 2.3.6.2 Search Space, Aggregation levels & common and UE specific
CCEs are numbered and used consecutively. A PDCCH with a format consisting of n CCEs may only
start with a CCE with a number equal to a multiple of n. Format is decided by eNB based on RF
conditions.
If a UE has good downlink RF (e.g. close to eNB), format 0 may be sufficient, but for a cell border UE,
format 3 may be required for robustness with good power level of a PDCCH.
1.3.7. PDCCH Candidate Selection

eNB should apply the following rules when allocating the PDCCH candidates. Let m = number of
PDCCH candidates to monitor in the given search space for all carriers and aggregation levels.
-
SI-RNTI / P-RNTI / RA-RNTI, use Common Search Space. UL/DL C-RNTI/ SPS C-RNTI, and DL
Temp. C-RNTI, use UE-Specific Search Space. TPC-PUCCH-RNTI / TPC-PUSCH-RNTI and UL
Temp. C-RNTI is not considered for default CCE management.
For SI-RNTI PDCCH candidate CCEs between 0 and (CS_Agr-1) is used and reserved in FDD and left
vacant if no SI-RNTI is scheduled. For TDD the default UL/DL configuration type 1, this PDCCH
candidate is reserved forS I-RNTI in sf 0 & 5 (and UL grant for C-RNTI/SPS-RNTI is not scheduled).
CCEs between CS_Agr and (2*CS_Agr-1) can be used either for P-RNTI or RA-RNTI.
For FDD:
-
For DL C-RNTI/SPS-RNTI/Temp C-RNTI the lowest m =m' from CCEs between 2*CS_Agr and
(Max_CCE-1) shall be used.
For UL C-RNTI/SPS-RNTI the lowest m =m">m' from CCEs between 2*CS_Agr and (Max_CCE-1)
shall be used.
For TDD:
- For DL C-RNTI/SPS-RNTI/Temp C-RNTI the lowest m =m' which has a PDCCH available from
CCEs between 2*CS_Agr and (Max_CCE-1) shall be used.
-
For UL C-RNTI/SPS-RNTI the lowest m =m">m' which from CCEs between 2*CS_Agr and
(Max_CCE-1) shall be used.
CCE resources utilized are well defined for default values of common search space aggregation level
=4, UE-specific search space aggregation L=2 resulting in 6 PDCCH candidates m=0..5. For different
Page 30
bandwidth, Max_CCE =20(5 Mhz)/25(10 MHz)/37(15 MHz)/50(20 MHz) for FDD. These are in general
to be applied in MAC Transport block size.
Each TDD subframe (take sf config 1 as example) having different PHICH group number, and for
5/10/15/20 MHz bandwidth, each subframe has, therefore, different number of MAX_CCE. SF0 and
SF5 cannot be used for UL grant. SF1 and SF6 are not used for DL assignment. SF2, SF3, SF7 and
SF8 are not applicable to PDCCH CCE allocation since they are uplink subframes.
1.3.8. Formats for Downlink Control Information (DCI)

The useful DCI content depends on the specific case of deployment. If no MIMO required, then MIMO
parameters are not required. To minimize overhead, different message formats are designed, each
with minimum payload required for a particular scenario.
The number of bits required for resource assignment depends on BW. To avoid complexity, Formats
0 and 1A are designed to be always 42bits. Since for different BW different size is needed, to reduce
complexity, the smaller format size is extended by adding padding bits to be the same size as the
larger. As an example, DCI message formats is listed below.
DCI
format
--------0
1
1A
1B
1C
1D
2
2A
3
3A
Bits for bandwidth of 50 RBs and

Purpose
four antennas at eNodeB)
--------------------------------------------------------------------------------------------------PUSCH grants
42
PDSCH assignments with a single codeword
47
PDSCH assignments using a compact format
42
PDSCH assignments for rank-1 transmission
46
PDSCH assignments using a very compact format
26
PDSCH assignments for multi-user MIMO
46
PDSCH assignments for closed-loop MIMO operation
62
PDSCH assignments for open-loop MIMO operation
58
Transmit Power Control (TPC) commands for multiple users
42
for PUCCH and PUSCH with 2-bit power adjustments
Transmit Power Control (TPC) commands for multiple users
42
for PUCCH and PUSCH with 1-bit power adjustments
Format 0. DCI Format 0 is used for resource grants for the PUSCH.
DCI Form ats D C I S T R Fields
Size
Format0
FreqHopping
1-bit
PUSCH frequency hopping flag
Allocation
variable
Resource block assignment/allocation
ModCoding
5-bits
Modulation, coding scheme and redundancy version
NewData
1-bit
New data indicator
TPC
2-bits
PUSCH TPC command
CShiftDMRS
3-bits
Cyclic shift for DM RS
CQIReq
1-bit
CQI request
TDDIndex
2-bits
For TDD config 0, this field is the Uplink Index.
DCIFormat
Description
Format0
For TDD Config 1-6, this field is the Dow nlink Assignment Index.
Not present for FDD
Page 31
Format 1. DCI Format 1 is used for resource assignments for single codeword PDSCH:
DCI Formats
DCISTR Fields
Format1
DCIFormat
Size
Description
Format1
AllocationType
1-bit
Resource allocation header: type 0, type 1

(only if downlink bandwidth is >10 PRBs)
Allocation
variable
ModCoding
5-bits
Modulation and coding scheme
HARQNo
3-bits (FDD)
HARQ process number
4-bits (TDD)
NewData
1-bit
New data indicator
RV
2-bits
Redundancy version
TPCPUCCH
2-bits
PUCCH TPC command
TDDIndex
2-bits
For TDD config 0, this field is not used.

For TDD Config 1-6, this field is the Downlink
Assignment Index. Not present for FDD
Format 1A. DCI Format 1A is used for compact resource assignments for single codeword PDSCH,
and allocating a dedicated preamble signature to a UE for contention-free random access:
DCI
Formats
Format1A
DCISTR Fields
Size
Description
DCIFormat
AllocationType
1-bit
Format1A
VRB assignment flag: 0 (localized), 1 (distributed)
Allocation
variable
ModCoding
5-bits
HARQNo
HARQ process number
NewData
3-bits (FDD)
4-bits (TDD)
1-bit
RV
2-bits
Redundancy version
TPCPUCCH
2-bits
PUCCH TPC command
TDDIndex
2-bits

Assignment Index.
New data indicator
Format 1B. DCI Format 1B is used for compact resource assignments for PDSCH using closed loop
precoding with rank-1 (transmission mode 6). Information is same as in Format 1A, but with addition
of precoding vector indicator applied for the PDSCH.
Format1B
DCIFormat
Format1B
AllocationType
1-bit
Allocation
variable
ModCoding
5-bits
HARQNo
3-bits (FDD)
HARQ process number
4-bits (TDD)
NewData
RV
1-bit
2-bits
New data indicator

Redundancy version
TPCPUCCH
2-bits
PUCCH TPC command
Page 32
TPMI
2-bits -(2 ants)
PMI information
4-bits- (4 ants)
PMI
TDDIndex
1-bit
2-bits
PMI confirmation
For TDD Config 1-6, this field is the Downlink Assignment Index.
Not present for FDD.
Format 1C. DCI Format 1C is used for very compact PDSCH assignments. With 1C format, PDSCH
is uses QPSK. This is used for paging, and some SI:
DCI
Formats
DCISTR Fields
Size
Description
Format1C
DCIFormat
Allocation
ModCoding
variable
5-bits
Format1C
Format 1D. DCI Format 1D is used for compact signalling of resource assignments for PDSCH using
multi-user MIMO (transmission mode 5). Information is similar as in Format 1B. Instead of one of
precoding vector indicators bits, there is a single bit for power offset indicators, to show if transmitted
power is shared between two UEs.
DCI Formats
DCISTR Fields
Format1D
DCIFormat
Size
-
Description
Format1D
AllocationType
1-bit
Allocation
variable
ModCoding
5-bits
HARQNo
3-bits (FDD)
HARQ process number
4-bits (TDD)
NewData
1-bit
New data indicator
RV
2-bits
Redundancy version
TPCPUCCH
2-bits
PUCCH TPC command
TPMI
2-bits (2 antenna)
Precoding TPMI information
4-bits (4 antenna)
DlPowerOffset
1-bit
Downlink power offset
TDDIndex
2-bits

Assignment Index.
Not present for FDD.
Format 2. DCI Format 2 is used for resource assignments for PDSCH for closed-loop MIMO
(transmission mode 4):
DCI
Formats
Format2
DCISTR Fields
DCIFormat
AllocationType
Size
1-bit
Description
Format2
Allocation
variable
TPCPUCCH
HARQNo
2-bits
3-bits (FDD)
4-bits (TDD)
1-bit
5-bits
PUCCH TPC command

HARQ process number
SwapFlag
ModCoding1
Transport block to codeword swap flag

Modulation and coding scheme for transport block 1
Page 33
NewData1
1-bit
New data indicator for transport block 1
RV1
ModCoding2
NewData2
RV2
2-bits
5-bits
1-bit
2-bits
Redundancy version for transport block 1

PrecodingInfo
3-bits - 2ants
6-bits - 4 ants
2-bits
Precoding information
TDDIndex

Assignment Index. Not present for FDD.
Format 2A. DCI Format 2A is used for resource assignments for PDSCH for open-loop MIMO
(transmission mode 3). Info is the same as Format 2, except that if eNodeB has two antenna ports,
there is no precoding information, and for four antenna ports two bits are used to indicate the
transmission rank.
DCI Formats
DCISTR Fields
Format2A
DCIFormat
Allocation Type
Size
1-bit
Description
Format2A
Allocation
variable
TPCPUCCH
2-bits
PUCCH TPC command
HARQNo
3-bits (FDD)
HARQ process number
4-bits (TDD)
SwapFlag
1-bit
Transport block to codeword swap flag
ModCoding1
5-bits
NewData1
1-bit
RV1
2-bits
ModCoding2
5-bits
NewData2
1-bit
RV2
2-bits
Precoding Info
0-bits
Precoding information
(2 antennas)
2-bits
(4 antennas)
TDDIndex
2-bits

Assignment Index.
Format 2B
DCI Formats
DCISTR Fields
DCIFormat
AllocationType
Size
1-bit
Description
Format2B
Format2B
Allocation
variable
TPCPUCCH
2-bits
PUCCH TPC command
HARQNo
3-bits (FDD)
HARQ process number
4-bits (TDD)
Page 34
ScramblingId
1-bit
Scrambling identity
ModCoding1
5-bits
NewData1
1-bit
RV1
2-bits
ModCoding2
5-bits
NewData2
1-bit
RV2
2-bits
TDDIndex
2-bits

Assignment Index.
Formats 3 and 3A. DCI Formats 3 and 3A are used for power control for PUCCH and PUSCH with 2bit or 1-bit power adjustments respectively.
DCI Formats
DCISTR Fields
Format3
DCIFormat
DCI Formats
Description
Format3
TPCCommands
variable
TPC commands for PUCCH and PUSCH
DCISTR Fields
Size
Description
DCIFormat
Format3A
Size
TPCCommands
variable
Format3A
TPC commands for PUCCH and PUSCH
CRC attachment. For UE to know whether it has received a PDCCH correctly, a 16-bit CRC is
appended to each PDCCH. CRC is scrambled with UE identity for this to be identified for a particular
UE. In UL MIMO, antenna may be indicated using Format 0 by antenna-specific mask to the CRC.
This way, no extra bit needed.
PDCCH construction. The PDCCH bits are encoded. The coded and rate-matched bits are then
scrambled with a cell-specific scrambling sequence to distinguish from neighbouring cells. The
scrambled bits are QPSK modulated and mapped to blocks of four REs (REGs). Interleaving is
applied for frequency diversity, followed by RE mapping to symbols indicated by PCFICH, excluding
PCFICH and PHICH. The PDCCHs are transmitted similar to PBCH, and diversity is applied if more
antenna ports are used.
1.3.9. Physical Hybrid ARQ Indicator Channel (PHICH)

PHICH carries the HARQ ACK(0)/NACK(1), indicating whether eNB has correctly received on
PUSCH. Bit is repeated in each of three BPSK symbols for robustness. Multiple PHICHs are mapped
to the same REs (of same PHICH group). Different PHICHs within group are separated through
different complex orthogonal Walsh sequences of length four for normal CP (two for extended CP).
Number of PHICHs in a group can be up to twice the sequence length. A cell-specific scrambling
sequence is then applied. PHICH duration (symbols) in time domain, is configurable by SI to either
one(normal) or three(extended) symbols.
Each of the three instances of orthogonal code of a PHICH is mapped to a REG on one of the first
three symbols of each subframe, such that each PHICH is partly transmitted on each available
symbols. UEs to deduce to which remaining resource elements in the control region the PDCCHs are
mapped.
Page 35
Fig 2.3.8.1 PHICH Coding, duration and cyclic shift allocation

PHICH index is implicitly associated with the index of the lowest uplink RB used for PUSCH. The
adjacent PUSCH RBs are associated with PHICHs in different PHICH groups, for load balancing.
Fig 2.3.8.2 PHICH bits to REG mapping

However, for MU-MIMO, this is not sufficient to enable multiple UEs to be allocated the same RBs for
a PUSCH. In this case, different cyclic shifts of RS are configured for different UEs for the same
PUSCH resources in time-frequency, and same cyclic shift index is then used to shift PHICH so that
each UE will receive its ACK or NACK on a different PHICH.
1.3.10. Resource Allocation

In each subframe, PDCCHs indicate resource allocations, normally localized, (Physical Resource
Block (PRB) in first slot is paired with PRB in the second slot of the subframe). Explanation here is in
terms of first slot only.
The most flexible/simple approach is to send each UE a bitmap, each bit indicating a particular PRB.
This is good for small BW, but for large BW (110 PRBs), bitmap would need 110 bits (too large). This
may make PDCCH larger than the data itself. One possible solution could have been to send a
combined message to all UEs, but that would need high power to ensure to reach each UE reliably.
Some methods are given below.
Page 36
Fig 2.3.9.1 Resource Allocation Type 0

Resource allocation Type 0. A bitmap indicates Resource Block Groups (RBGs) allocated to
scheduled UE, where a RBG is a set of consecutive PRBs. RBG size (P=1,2,3,4) depends on BW.
Total number of RBGs = NRBG = Total_RB /P.
Resource allocation Type 1. Here Individual PRBs can be addressed only within a subset of the
PRBs available. Bitmap is slightly smaller than Type 0, since some bits are used to indicate the
subset of addressed RBG, and a shift in position. Total bits is still the same as for Type 0. One bit is
used for subset selection and another bit to indicate the shift. The provides flexibility in spreading the
RBs across BW to exploit frequency diversity.
Fig 2.3.9.2 Resource Allocation Type 1
Resource allocation Type 2. Resource allocation information indicates to a scheduled UE either:

a. a set of contiguously allocated PRBs, or
b. a distributed allocation comprising multiple non-consecutive PRBs.
A 1-bit flag in Resource allocation message indicates Contigous or Distributed. PRB allocations may
vary from single PRB up to maximum number of PRBs spanning BW. Type 2 allocation field has
Page 37
Resource Indication Value (RIV) corresponding to a starting RB(RB START) and a length in terms of
contiguously-allocated resource blocks (LCRBs). RIV is defined by:
if (LCRBs 1) NDLRB /2 then RIV = NDL,RB (LCRBs 1) + RBSTART
(start from lower)
else RIV = NDL,RB (NDL,RB LCRBs + 1) + (NDL,RB 1 RBSTART)
Example: for Localized mode, RBStart = 5, LCRB = 20 and NDL,RB = 50
Here: RIV = 50 * (20-1) + 5 = 955
1.3.11. DL Resource Allocation Rules

The DL resource allocation is an eNB function. When the DL data is to be sent with a specific
scheduling requirement, for instance, in a TTI in advance rather than now, This advance time in
general covers all time delays. Two types of DCI combinations are identified as default formats for the
signalling and protocol test.
1. DCI combination 1 uses:
a. DCI format 1A, resource allocation type 2 localised, for all DL scheduling types.
2. DCI combination 2 uses:
a. DCI format 1C, resource allocation type 2 distributed, for scheduling of PCCH/BCCH/RAR;
and
b. DCI format 1 resource allocation type 0, for UE dedicated scheduling.
Usually Transmission mode 3 is used in MIMO test cases with 2 Transmit antenna, where UE is
expected to decode only DCI formats 2A and 1A. Similarly for TM4, UE is expected to decode only
DCI formats 2 and 1A.
General DL scheduling scheme
The bandwidth of 5/10/20 MHz makes 25/50/100 available physical resource blocks respectively.
These resource blocks are divided into three distinct sets. Exact set sizes and the elements contained
in the individual sets depend upon the DCI combination to be applied.
The first set is reserved for BCCH mapped to DL-SCH (SI-RNTI).
The second set is reserved for PCCH mapped to DL-SCH (P-RNTI).
The third set is used for one of mutually exclusive transmissions of:
a. 'Random Access Response' mapped to DL-SCH (RA-RNTI); or
b. UE-dedicated scheduling mapped to DL-SCH (C-RNTI/ SPS C-RNTI/ Temp C-RNTI).
For each subframe where data is scheduled, eNB shall select a Transport Block Size (TBS),
independently for each type of data scheduled, such that:
All the scheduled data is transmitted respecting the timing information.
Not more than MaxRbCnt resource blocks are used, for DCI format 1C, NPRB = MaxRbCnt.
Minimum MAC Padding is performed.
If all scheduled Data cannot be transmitted in the indicated subframe, for example due to TDD and half
duplex configuration, it shall be transmitted in the next available subframe.
Additional rules for BCCH scheduling scheme
This scheme is applicable for Data transmission on logical channel BCCH mapped to DL-SCH,
PDCCH scrambled by SI-RNTI. 4 physical resource blocks are reserved for BCCH transmission with
QPSK. Following additional rules are applied for TBS selection:
- The Max TBS, the maximum TBS allowed for the scheduling scheme, is restricted to 600. (nearest value
achievable for ITBS = 9 and NPRB = 4.
- If BCCH cannot fit into a TBS smaller or equal to Max TBS, its an error.
Additional rules for PCCH specific scheduling scheme
PDCCH scrambled by P-RNTI. For DCI combination 1, one physical resource block is reserved for
this DL-SCH. For DCI combination 2, two PRBs are reserved for 5 MHz bandwidth, and four PRBs
are reserved for 10 MHz or 20 MHz bandwidth with QPSK.
Page 38
Additional rules for RAR specific scheduling scheme
For transmission of Random Access Response mapped to DL-SCH, PDCCH is scrambled by RARNTI. four PRBs are reserved with QPSK. The Max TBS is restricted to 600 bits (nearest value
achievable for ITBS = 9 and NPRB =4.
Additional rules for UE-dedicated scheduling scheme in normal mode
For DL subframe UE-dedicated resources, these are mapped to DL-SCH, PDCCH scrambled by CRNTI/ SPS C-RNTI/ Temp C-RNTI. Maximum modulation is restricted to 64QAM. For DCI
combination 1, 20 PRBs (5 to 24), and for DCI combination 2, 17 PRBs are reserved.
In TDD no data is transmitted in DwPTS of the special subframe.
HARQ Retransmission Resource
The eNB should support DL QPSK, 16QAM and 64QAM modulation schemes. The configured radio
parameters, including DCI format, resource allocation types, maximum allowed modulation scheme,
first virtual / physical resource block to be used, maximum available RBs and RV are all known to
eNB.
If in a TTI more than one transport blocks are scheduled (DCI format 2/ 2A/2B), HARQ retransmission
is handled independently for each TB by eNB. In case UE ACKs one TB and NACKs the other and
there is no fresh data scheduled for transmission, eNB only schedules the NACKed TB for
retransmission, using same Imcs as used in initial transmission, mapped to codeword 0. Acked TB
(and hence codeword 1) is disabled by setting corresponding I MCS 0 and rvidx = 1. Resource
allocation (Nprb) used in retransmission is same as in initial transmission.
1.3.12. Resource Allocation Bitmap examples

Example: Physical resource allocation bitmap for DCI combination 1 (5/10/20
MHz)
Example: Physical resource allocation bitmap for DCI combination 2 (5/10/20

MHz)
Page 39
1.3.13. Uplink Grant

The Network informs the UE if it is allowed to make Uplink Data transmission by transmitting 'DCI
format 0' on PDCCH. UE shall transmit (4 TTI later for FDD or variable for TDD) a TB of exactly the
same size as specified in DCI format 0. The UE has no control of its own on TB size, and has to
merely follow the network, even if that means lots of MAC padding or resource starving. The UE has
the following means to communicate if it has UL data ready for transmission and subsequently the
estimate of quantity of data to be transmitted.
RACH procedure: UE in idle mode, handed over to a new cell or connected mode but PUCCH is
unsynchronized (sometimes referred to as PUCCH is not configured) will trigger RACH procedure on
data ready for transmission in UL.
Scheduling Request: UE in connected mode, no grant configured, PUCCH is synchronized and

has data ready for transmission in UL, will transmit a scheduling request on PUCCH.
Buffer Status Reports: UE in connected mode, PUCCH synchronized, has a configured grant
for current TTI, but grant is not sufficient to transmit all the data will include MAC control element BSR
in the UL MAC PDU.
RACH and SR indicate on data availability and BSR provides an estimate of data available for
transmission. CQI/PMI/RI feedback from the UE which indicates the channel conditions and
recommended number of layers.
Hence to determine the exact need of the grant requirement of the UE a network needs to act on all
four of the above.
The NW disables aperiodic CQI/PMI/RI feedback from the UE by setting the CQI request field to 0 in
DCI format 0/RAR grant.
eNB, will periodically transmit automatically MAC PDUs containing the MAC control element 'Timing
Advance'. The period normally is set to 80 % of the 'Time Alignment Timer' default value (750 ms)
configured at UE.
Page 40
Network will UL grant at every reception of a Scheduling Request. All UL grant define grant allocation
in terms of IMCS and NPRB to be used. The eNB shall allocate RBs corresponding to PRB indices
0..(NPRB-1).
1.3.14. PDCCH Transmission and Blind Decoding

How PDCCH transmissions are organized so that a UE can locate PDCCHs intended for it?
One approach can be that, eNodeB should be able to place any PDCCH anywhere by CCEs area
indicated by PCFICH. UE would need to check all possible PDCCH locations, PDCCH formats and
DCI formats, check CRCs (scrambled with UE identity). Carrying out blind decoding of all possible
combinations would require many PDCCH decoding attempts in every subframe. For large BW, it
would become a significant burden. For example, blind decoding of 100 possible CCE locations for
PDCCH Format 0 would be equivalent to continuously receiving a data rate of around 4 Mbps.
The alternative approach is to define for each UE a limited set of CCE locations where a PDCCH may
be placed. Set of such CCE locations can be considered as a search space for that UE. Search
space has a different size for each PDCCH format. Moreover, separate dedicated and common
search spaces are defined, where a dedicated search space is configured for each UE individually,
while all UEs are informed of the extent of the common search space. Dedicated and common search
spaces may overlap for a given UE.
With such small search spaces, eNodeB may not find CCE to send PDCCHs to all the UEs. To
minimize blocking persisting into next subframe, a UE-specific hopping sequence is applied to the
starting positions of the dedicated search spaces.
To keep blind decoding attempt computational load, UE is not required to search for all DCI formats
simultaneously. Typically, in dedicated search space, UE will always search for Formats 0 and 1A. In
addition, UE may receive a further format 1, 1B or 2 DCI depending on PDSCH transmission mode.
In the common search space, UE will search for Formats 1A and 1C. In addition UE may be
configured to search 3 or 3A DCI, and may be distinguished by CRC scrambled by a different
(common) identity, rather than a UE-specific one. This way not it is limited blind decoding required.
1.3.15. Enhanced PDCCH (EPDCCH)

-
Informs the UE about the resource allocation of DL-SCH, and Hybrid ARQ information related to
DL-SCH;
- Carries the uplink scheduling grant.
The EPDCCH carries UE-specific signalling. It is located in UE-specifically configured PRBs and
consists of:
- Transport format, resource allocation, and HARQ# for DL-SCH;
- Transport format, resource allocation, and HARQ# for UL-SCH;
Multiple EPDCCHs are supported and a UE monitors a set of EPDCCHs.
EPDCCHs are formed by aggregation of enhanced control channel elements (ECCEs), each eCCE
consisting of a set of resource elements. Different code rates for EPDCCHs are realized by
aggregating different numbers of eCCEs. An EPDCCH can use either localized or distributed
transmission, differing in the mapping of enhanced control channel elements to the resource elements
in the PRBs.
EPDCCH supports C-RNTI and SPS C-RNTI. If configured, EPDCCH is applicable in the same way
as PDCCH unless otherwise specified.
EPDCCH assignment procedure

For each serving cell, UE can be configured with one or two EPDCCH-PRB-sets for EPDCCH
monitoring. The PRB-pairs are indicated. Each EPDCCH-PRB-set consists of set of ECCEs
numbered from 0 to
N ECCE, p ,k 1 where N ECCE, p ,k is the number of ECCEs in EPDCCH-PRB-set p
of subframe k . Each EPDCCH-PRB-set can be configured for either localized EPDCCH transmission
or distributed EPDCCH transmission.
Page 41
UE shall monitor a set of EPDCCH candidates on one or more activated serving cells as configured in
UE-specific search spaces. For each serving cell, the subframes in which the UE monitors EPDCCH
UE-specific search spaces are configured by search patterns.
If the UE is configured with a carrier indicator field, then the UE shall monitor one or more EPDCCH
UE-specific search spaces at each of the aggregation levels on one or more activated serving cells.
On each serving cell c, UE shall monitor EPDCCH with CRC scrambled by C-RNTI and SPS C-RNTI
in the EPDCCH UE specific search space of serving cell c.
A UE is not expected to monitor the EPDCCH of a secondary cell if it is configured to monitor
EPDCCH with carrier indicator field corresponding to that secondary cell in another serving cell. For
the serving cell on which EPDCCH is monitored, the UE shall monitor EPDCCH candidates at least
for the same serving cell.
EPDCCH starting position

For a given Scell, for tm 1-9,
if the UE is configured with epdcch-StartSymbol-r11,
o starting OFDM symbol for EPDCCH = l EPDCCHStart in the first slot in a subframe,
otherwise
o starting OFDM symbol for EPDCCH =
o
l EPDCCHStart in the first slot in a subframe = CFI
value
for each EPDCCH-PRB-set, starting OFDM symbol for monitoring EPDCCH in
subframe k is determined from PDSCH starting position for PDSCH RE mapping as
follows
if the value of PDSCH starting position for PDSCH RE mapping is 5,
l' EPDCCHStart = CFI value
otherwise
l' EPDCCHStart =PDSCH starting position for PDSCH RE mapping
if subframe k is MBSFN subframe configuration for PDSCH RE mapping

l EPDCCHStart min( 2, l ' EPDCCHStart ) ,
otherwise
'
lEPDCCHStart lEPDCCHStar
.
t
Mapping to resource elements

ePDCCH symbols are mapped in sequence on the associated antenna port when:
- they are part of the EREGs assigned for the EPDCCH, and
- they are not part of a PBCH or synchronization signals, and
- they are not part of CRS, CSI RS for the specific UE, and
- index l in the first slot in a subframe fulfils l lEPDCCHStart .
The mapping to resource elements k, l on antenna port p shall be in increasing order of first k and
then l , starting with the first slot and ending with the second slot in a subframe.
For localized transmission, the single antenna port p to use is given by
ECCE
ECCE
ECCE
n' nECCE,low mod N RB
nRNTI mod min(N EPDCCH
, N RB
)
where nECCE,low is the lowest ECCE index used by this EPDCCH transmission in the EPDCCH set,
ECCE
nRNTI corresponds to the RNTI associated with the EPDCCH transmission, and N EPDCCH
is the
number of ECCEs used for this EPDCCH.

The demodulation reference signal associated with EPDCCH
- is transmitted on the same antenna port p 107,108,109,110 as the associated EPDCCH
physical resource;
Page 42
is present and is a valid reference for EPDCCH demodulation only if the EPDCCH transmission
is associated with the corresponding antenna port;
- is transmitted only on the physical resource blocks upon which the corresponding EPDCCH is
mapped.
A demodulation reference signal associated with EPDCCH is not transmitted in resource elements
k, l in which one of the physical channels or physical signals other than the demodulation reference
signals are transmitted using resource elements with the same index pair k, l regardless of their
antenna port p .
Resource mapping parameters for EPDCCH

For an Scell, if the UE is configured to receive PDSCH data in tm10, and if the UE is configured to
monitor EPDCCH, for each EPDCCH-PRB-set, the UE shall use the parameter set indicated by reMappingQCLConfigListId-r11 for determining the EPDCCH RE mapping and EPDCCH antenna port
quasi co-location. The following parameters are included in the parameter set:
Number of CRS antenna ports for PDSCH RE mapping.
CRS frequency shift for PDSCH RE mapping.
MBSFN subframe configuration for PDSCH RE mapping.
Zero-power CSI-RS resource configuration(s) for PDSCH RE mapping.
PDSCH starting position for PDSCH RE mapping.
CSI-RS resource configuration identity for PDSCH RE mapping.

For a given serving cell, for each EPDCCH-PRB-pair set
p , the UE is configured with
resourceBlockAssignment-r11 indicating a combinatorial index r .
EPDCCH formats
The EPDCCH carries scheduling assignments which is transmitted using an aggregation of one or
several consecutive enhanced control channel elements (ECCEs) where each ECCE consists of
multiple enhanced resource element groups (EREGs). The number of ECCEs used for one EPDCCH
depends on the EPDCCH format and the number of EREGs per ECCE is pre-defined. Both localized
and distributed transmission is supported.
An EPDCCH can use either localized or distributed transmission, differing in the mapping of ECCEs
EREG
to EREGs ( N ECCE
= 4 for normal or 8 for extended CP per ECCE) and PRB pairs.
A UE shall monitor multiple EPDCCHs. One or two sets of PRB pairs which a UE shall monitor for
EPDCCH transmissions can be configured. All EPDCCH candidates in EPDCCH set S m use either
only localized or only distributed transmission as configured. Within EPDCCH set S m in subframe i ,
the ECCEs available for transmission of EPDCCHs are numbered from 0 to N ECCE,m,i 1 and ECCE
number n .
EREG
ECCE
EREG
N ECCE
16 N ECCE
is the number of EREGs per ECCE, and N RB
is the number of ECCEs/RB pair.
The PRB pairs constituting EPDCCH set S m are assumed to be numbered in ascending order from 0
Sm
to N RB
1.
DL
25 , - 2,4,8,16 or 32 ECCEs may
When DCI formats 2, 2A, 2B, 2C or 2D is used and N RB
be there
- any DCI format when nEPDCCH 104 and normal cyclic prefix is used in normal subframes or
special subframes with configuration 3, 4, 8(TDD) - 2,4,8,16 or 32 ECCEs may be there
- otherwise - 2,4,8,16 or 32 ECCEs may be there
The quantity nEPDCCH for a particular UE is defined as the number of downlink resource elements
(k , l ) in a PRB pair configured for possible EPDCCH transmission of EPDCCH set S 0 and fulfilling all
of the following criteria:
- they are part of any one of the 16 EREGs in the PRB pair, and
Page 43
they are assumed by the UE not to be used for CRS or CSI RS, and
the index l in the first slot in a subframe fulfils l lEPDCCHStart .
EPDCCH-Config
The IE EPDCCH-Config is used to configure the subframes and resource blocks for EPDCCH
monitoring.
-- ASN1START
EPDCCH-Config-r11 ::=
SEQUENCE{
epdcch-SubframePatternConfig-r11 CHOICE {
release
NULL,
setup
SEQUENCE {
epdcch-SubframePattern-r11
MeasSubframePattern-r10
}
}
epdcch-StartSymbol-r11
INTEGER (1..4)
epdcch-SetConfigReleaseList-r11
EPDCCH-SetConfigReleaseList-r11
epdcch-SetConfigAddModList-r11
EPDCCH-SetConfigAddModList-r11
}
OPTIONAL, -- Need ON
OPTIONAL, -- Need OP
OPTIONAL, -- Need ON
OPTIONAL -- Need ON
EPDCCH-SetConfigAddModList-r11 ::=
SEQUENCE (SIZE(1..2)) OF EPDCCH-SetConfig-r11
EPDCCH-SetConfigReleaseList-r11 ::=
SEQUENCE (SIZE(1..2)) OF EPDCCH-SetIdentity-r11
EPDCCH-SetConfig-r11 ::=
SEQUENCE {
epdcch-SetIdentity-r11
epdcch-TransmissionType-r11
epdcch-ResourceBlockAssignment-r11
numberPRBPairs-r11
resourceBlockAssignment-r11
},
dmrs-ScramblingSequenceInt-r11
pucch-ResourceStartOffset-r11
re-MappingQCLConfigListId-r11
}
EPDCCH-SetIdentity-r11 ::=
EPDCCH-SetIdentity-r11,
ENUMERATED {localised, distributed},
SEQUENCE{
ENUMERATED {n2, n4, n8},
BIT STRING (SIZE(4..38))
INTEGER (0..503),
INTEGER (0..2047),
PDSCH-RE-MappingQCL-ConfigId-r11
OPTIONAL -- Need OR
INTEGER (0..1)
-- ASN1STOP
EPDCCH
dmrs-ScramblingSequenceInt - The DMRS scrambling sequence initialization parameter nID,i
.
epdcch-SetConfig - Provides EPDCCH configuration set. E-UTRAN configures at least one epdcchSetConfig when EPDCCH-Config is configured.
epdcch-SetIdentity - Indicates the indentity of the EPDCCH set.
epdcch-StartSymbol (1,2,3,4) - Indicates the OFDM starting symbol for any EPDCCH and PDSCH
scheduled by EPDCCH on the same cell, if the UE is not configured with tm10. If not present, the
configuration is released and the UE shall derive it from PCFICH. It is not configured for UEs
configured with tm10.
epdcch-SubframePatternConfig - Configures the subframes which the UE shall monitor the UEspecific search space on EPDCCH. If it is not configured when EPDCCH is configured, the UE
monitors the UE-specific search space on EPDCCH in all subframes except for pre-defined rules.
epdcch-TransmissionType - Indicates whether distributed or localized EPDCCH transmission mode
is used.
numberPRBPairs - Indicates the number of PRB pairs used for the EPDCCH set. Value n2
corresponds to 2 PRB pairs; n4 corresponds to 4 PRB pairs and so on. n8 is not supported for dlBandwidth having value n6.
pucch-ResourceStartOffset - PUCCH format 1a and 1b resource starting offset for the EPDCCH
set.
re-MappingQCLConfigListId - Indicates the starting OFDM symbol, the related rate matching
parameters and quasi-collocation assumption for EPDCCH when the UE is configured in tm10. This
Page 44
provides the index of PDSCH-RE-MappingQCL-ConfigId. E-UTRAN configures this only when tm10 is
configured.
resourceBlockAssignment - Indicates the index to a specific combination of PRB pair for EPDCCH
set. The size of resourceBlockAssignment is calculated based on numberPRBPairs and the signalled
value of dl-Bandwidth.
1.3.16. Scheduling Process - Control Channel Viewpoint

To summarize operation of PDCCH, sequence of steps carried out by eNodeB are:
1. Determine UEs, who should be granted UL resources, based on CQI, SR and BSR.
2. Determine UEs, who should be scheduled for DL, based on CQI, RI and PMI.
3. Identify any common control messages required (e.g. power control using DCI Format 3).
4. For each message decide PDCCH format (i.e. 1, 2, 4 or 8 CCEs), power offset to reach the
UE(s) with sufficient reliability, while minimizing PDCCH overhead.
5. Determine how many CCEs required, symbols needed PDCCHs and compute PCFICH.
6. Map each PDCCH to a CCE location within appropriate search space.
7. If any PDCCHs cannot be mapped to a CCE (congested), either:
a. continue to next step (step 8) accepting likely loss in throughput, or:
b. allocate one more symbol to support required PDCCHs by PCFICH and possibly
revisit step 1 and/or 2 and change UE selection and allocation (to fully use UL & DL
resources).
8. Allocate necessary resources to PCFICH and PHICH.
9. Allocate resources to PDCCHs.
10. Check total power level per symbol does not exceed maximum, and adjust if necessary.
11. Transmit downlink control channels- PCFICH, PHICH and PDCCH.
Page 45
Page 46
1.4 Physical Layer Processing - DL

LTE adjusts (adapts) the data rate (modulation & coding rate scheme (MCS)) dynamically to match
the prevailing radio channel capacity for each user.
The eNodeB selects modulation & code rate depending on a prediction of DL channel conditions,
based on Channel Quality Indicator (CQI) transmitted by the User Equipment (UE) in UL. CQI is an
indication of data rate which can be supported by the channel, based on SINR and UEs receiver. The
eNodeB can select different CQI feedback modes to trade off the improved DL link adaptation
enabled by CQI against UL overhead caused by the CQI itself.
Signalling necessary for interoperability between eNodeB and UEs are defined for link adaptation, but
exact methods used by eNodeB are left to the manufacturers choice. In response to CQI, eNodeB
can select between QPSK, 16-QAM and 64-QAM schemes and a range of code rates.
Overall Channel Coding Chain for Data
The physical layer first attaches a 24-bit CRC to each TB received from MAC layer. This is used by
receiver to verify correct reception and to generate ACK/NACK. The TB is then segmented into code
blocks designed to minimize filler bits needed to match the available QPP sizes. Two adjacent QPP
sizes are allowed for segmenting a TB, rather than being restricted to a single QPP size. Filler bits
would then be placed in the first segment. LTE specification always recommends single QPP size for
each segment.
Following segmentation, a further 24-bit CRC is attached to each code block if the TB was split into
two or more code blocks. Note that, polynomial used for the code-block-level CRC is different from
the polynomial used for the transport block CRC, done deliberately to avoid increasing the probability
of failing to detect errors as a result of the use of individual CRCs per code block; if all the code block
CRCs pass, decoder should still check the transport block CRC, is likely to detect an error which was
not detected by a code-block CRC.
Although code block concatenation is done before scrambler and modulation mapper, each code
block is associated with a distinct set of modulation symbols, implying that scrambling and modulation
mapping operations may be done individually for each code block (for efficiency).
a0 , a1 ,..., a A 1
Transport block
CRC attachment
b0 , b1 ,..., b B 1
Code block segmentation

Code block CRC attachment
cr 0 , cr1 ,..., cr K r 1
Channel coding
d r(i0) , d r(1i ) ,..., d r(i)D
r 1
Rate matching
er 0 , er1 ,..., er Er 1
Code block
concatenation
f 0 , f1 ,..., f G 1
Fig: 2.4.0 DL-PDSCH Channel Processing
Page 47
1.4.1. Link Adaptation and Feedback Computation

Quality of signal received by UE depends on channel quality, interference and noise level from
serving cell. To optimize capacity and coverage. The eNB transmitter should try to match data rate for
each user to the variations in received quality. This is link adaptation and is based on Adaptive
Modulation and Coding (AMC), consisting of modulation and coding schemes:
Modulation Scheme. Low-order modulation (lower bit rate) is more robust and can tolerate
higher levels of interference and vice versa. Higher modulation is useful only when the SINR is
sufficiently high.
Code rate. For a given modulation, code rate is chosen depending on the radio link conditions:
lower code rate in poor SINR and a higher code rate in high SINR. Code rate adaptation is
achieved by puncturing or repetition to the output of a mother code.
Should all RBs allocated to one UE in a subframe use same MCS or should be frequency dependent?
In general only a small throughput improvement arises from a frequency-dependent MCS compared
to an RB-common MCS, hence frequency-dependent MCS is not justified, and they are constant over
the allocated RB for a given user, and time-domain scheduling and AMC is supported. If multiple TBs
are transmitted to one user in a subframe using MIMO, each TB can use an independent MCS.
CQIs assists the eNodeB in selecting MCS for DL. CQI is derived from DL RSRQ of Ref Signal,
Reported CQI indicates highest MCS it can decode with a TB error rate probability not exceeding
10%, and not the SINR. Hence a UE designed with better algorithms (like interference cancellation
techniques) can report higher CQI and can receive a higher data rate. CQI could be based on a set of
1
Block Error Rate (BLER) thresholds. UE reports CQI value for MCS to ensures BLER 10 based on
measured received signal quality.
1.4.2. CQI Feedback in LTE

CQI report is controlled by eNodeB, both periodic and aperiodic. PUCCHis used for periodic CQI only;
PUSCH is used for aperiodic CQI, whereby eNodeB specifically instructs UE to send CQI embedded
into a resource scheduled for UL.
Frequency granularity of CQI is determined by subbands (N), each comprised of k contiguous PRBs.
N = _NRB /k. The CQI reporting modes can be Wideband CQI, eNodeBconfigured sub-band feedback,
or UE-selected sub-band feedback. In MIMO in eNodeB, CQI value(s) may be reported for a second
codeword.
Aperiodic CQI Reporting
Aperiodic CQI on PUSCH is scheduled by setting CQI request bit in UL grant on PDCCH.
The CQI reporting type can be:
Wideband feedback
UE reports one wideband CQI value for the whole BW.
eNodeB-configured sub-band feedback
UE reports a wideband CQI value for the whole system bandwidth. Also, UE reports a CQI value
for each subband, calculated assuming transmission only in the relevant sub-band. Sub-band CQI
reports are encoded using 2-bits as follows:
Sub-band differential CQI offset = Sub-band CQI index Wideband CQI index Possible sub-band
differential CQI offsets are { 1, 0, +1,+2}.
UE-selected sub-band feedback
UE selects M preferred sub-bands of size k within whole BW. UE reports one wideband CQI and
one CQI value reflecting average quality of M selected sub-bands, including positions of M
selected sub-bands. The CQI value for the M selected sub-bands for each codeword is encoded
differentially using 2-bits relative to its respective wideband CQI as defined by:
Differential CQI = Index for average of M preferred sub-bandsWideband CQI index
Periodic CQI Reporting
If eNodeB wishes to receive periodic reporting of the CQI, UE will transmit using the PUCCH. Only
wideband and UE-selected sub-band feedback is sent for periodic CQI reporting. With aperiodic CQI
reporting, it is configured by eNodeB by RRC. For wideband periodic CQI, the period can be {2, 5, 10,
16, 20, 32, 40, 64, 80, 160} ms or Off.
Page 48
Wideband feedback mode is similar to via PUSCH, the UEselected sub-band CQI using PUCCH is
different. In this case, number of sub-bands N is divided into J bandwidth parts. The J depends on
BW.
In periodic UE-selected sub-band CQI reporting, one CQI value is computed and reported for a single
selected sub-band from each bandwidth part, along with the corresponding sub-band index.
1.4.3. Channel Coding

Theoretical Aspects of Channel Coding
From Convolutional Codes to Turbo Codes
Convolutional encoders are for small data blocks and Turbo encoder are for relatively bigger data
blocks. An encoder is represented as C(k, n, m), where
It has m stages shift registers.
At each instant, k bits enter registers and k bits in last position of shift register are
dropped.
Set of n output bits is a linear combination of the content of the shift register.
The Code rate is defined as Rc = k/n.
Example LTE 1/3 Convolutional encoder has
o m = 6, n = 3, k = 1 and rate Rc = 1/3.
o N=3 generator sequences are represented by octets g0=133, g1=171,
g2=165
Trellis diagram represents a finite state machine including time dimension. Consider an input block
with L bits encoded with a rate 1/n (i.e. k = 1) convolutional encoder, resulting in a codeword of length
(L + m) n bits, including m trellis termination bits (or tail bits) inserted at the end of the block to drive
the shift register contents back to all zeros at the end of the encoding process.Using tail bits is just
one possible way of terminating an input sequence.
Tail Biting Method: Method of simple truncation (no tail bits appended) is called tail-biting. In tailbiting, initial and final states of convolutional encoder are required to be identical. Usually tail-biting for
feed-forward convolutional encoders is achieved by initializing the shift register contents with the last
m bits in the input block. Tailbiting encoding facilitates uniform protection of the information bits and
suffers no rate-loss owing to the tail bits. Tail- can be decoded using, Circular Viterbi Algorithm (CVA).
1.4.4. Viterbi Algorithm (VA) (Example):

Lets take a simple convolutional code with generator polynomials g0 = [133], g1 = [171] and g2 =
7[165]. Each edge in the trellis corresponds to a transition from a state s(all zero) to a state s(all zero),
Shift registers are initialized to all-0 state and m tail bits are added at the end (all zeros).
ck
d k(0) G0 = 133 (octal)
d k(1) G1 = 171 (octal)

d k( 2) G2 = 165 (octal)
Fig 2.4.4.1 Sample Convolutional Encoder

Let M(yi | xi) = (j..n)( log P(yi,j | xi,j )) be branch metric at the ith trellis step. The VA, using trellis,
computes best partial path metric at each step by adding, comparing and selecting metrics. For each
state s, VA computes the possible partial path metrics corresponding to all the edges arriving in state
s, and selects the best partial metric. In this example, at time = 2, there are three possible paths
ending. The VA computes and selects the best path metrics as the survival edge and the other one is
discarded. This is carried on for each state and for = 0, . . . , L + m 1. At the last stage l=L+m1, VA
Page 49
selects the best total metric among different metrics computed for each state (in case of trellis
terminated code, last state is all-0 state) and it traces back the selected path in the trellis to provide
the estimated input sequence. This is a recursive iterative process.
Convolutional codes are most widely used for good performance, decoding speed based on VA and
flexible codeword sizes adaptation. Followed by this, turbo code and Low-Density Parity Check
(LDPC) codes were discovered that provided near-Shannon limit performance.
Turbo Codes
Turbo codes is an iterative decoding algorithm to achieve near-Shannon limit performance. Encoder
has two convolutional encoders linked by an interleaver. Two identical convolutional codes have g0 =
[13] and g1 = [15]. Turbo code encodes the input block twice (with and without interleaving) to
generate two set of parity bits. Each encoder is terminated to all zero state by using tail bits. The
nominal code rate of turbo code is 1/3.
Number of states in the trellis of a turbo code is significantly larger due to the interleaver, making it
intractable (except for trivial block sizes). Therefore, iterative decoding is done based on separate
optimal decoder for each constituent convolutional coder, both iteratively exchanging bits via a
(de)interleaver.
The two decoders cooperate by iteratively exchanging bits via (de)interleaver. After a certain number
of iterations, the output can be used to obtain final hard decision estimates of the information bits.
xk
1st constituent encoder
ck
zk
Output
Input
Turbo code internal

interleaver
2nd constituent encoder
zk
Output
ck
xk
Fig: 2.4.4.2 Turbo Encode

Channel Coding for Data Channels in LTE
Turbo codes were immediately used in UMTS and now they are further enhanced.
Channel coding
UMTS
LTE
Constituent code
Tailed, eight states, R = 1/3
Same
mother code
Turbo interleaver
Row/column permutation
Contention-free quadratic
permutation polynomial (QPP)
interleaver
Rate matching
Hybrid ARQ
Performed on concatenated
Virtual Circular Buffer (CB) rate
code blocks
matching, performed per code block
Redundancy Versions (RVs)
RVs defined on virtual CB, Chase
Page 50
defined, Chase operation
operation allowed
Allowed
Control channel
256-state tailed
64-state tail-biting convolutional
convolutional code
code, CB rate matching
Turbo coding only
CRC attachment, turbo coding, rate
Per code block

operations
matching, modulation mapping
It was enhanced by the ability to select different redundancy versions for HARQ retransmissions.
However the decoder shows the strain at 10Mbps. LTE effort began for data rates of 100 Mbps to
1Gbps in view. For LTE, turbo interleaver was replaced with a contention-free interleaver.
Contention-Free Turbo Decoding
The existing UMTS interleaver had a problem with memory access contentions (read or write from/to
the same memory same time. Contention resolution is possible with extra hardware, and the
resolution time (cycles) may vary for every interleaver size. Complex memory management is used as
contention resolution for any arbitrary interleaver, such that no contentions occur.
It requires that for each window, the memory banks accessed be unique between any two windows,
thus eliminating access contentions. Instead of using M separate memories, better to use single
physical memory and fetch/store M values on each cycle from a single address. This requires CF
interleaver to satisfy a vectorized decoding property where the intra-window permutation is the same
for each window.
A variety of possible parallelism factors provides freedom for each individual manufacturer to select
the degree of parallelism based on the target data rates for different UE categories. After
consideration of performance, available flexible classes of CF interleavers and complexity benefits, a
new contention-free interleaver was selected for LTE.
1.4.5. LTE Contention-Free Interleaver

2
For block size K, a QPP(Quadratic Permutation Polynomial) interleaver is defined by: (i) = (f1i + f2i )
mod K, where
o i is the output index (0 i K 1),
o (i) is the input index and
o f1 and f2 are the coefficients:
f1 is relatively prime to block size K;
all prime factors of K also factor f2.
i
f1
f2
f1
f2
f1
f2
f1
f2
1
2
-46
47
40
48
--400
408
3
7
--151
155
10
12
--40
102
48
49
--93
94
416
424
--1056
1088
25
51
--17
171
52
106
--66
204
95
96
--140
141
1120
1152
--3072
3136
67
35
--47
13
140
72
--96
28
142
143
--187
188
3200
3264
--6080
6144
111
443
--47
263
240
204
--190
480
Fig 2.4.5 Interleaver position converter f1-f2 Table

A total of 188 interleavers are defined for LTE, of which 153 have quadratic inverses while the
remaining 35 have degree-3 and degree-4 inverses. Attractive feature of QPP interleavers is that they
are maximum contention-free, supporting parallelism. For example, for K = 1024, supported
parallelism factors include {1, 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024}, although factors that result in a
window size less than 32 may not be required in practice.
QPP interleavers also have even-even property whereby even and odd indices in the input are
mapped to even and odd indices respectively in the output; this enables the encoder and decoder to
process two bits per clock cycle.
Page 51
The amounts of parallelism depend on the factorization of the block size, certain block sizes (prime
sizes) are not natively supportable by the turbo code. Filler bits are used to pad the input to the
nearest QPP interleaver size. The QPP sizes are selected such that:
Number of interleavers is limited (fewer interleavers implies more filler bits).
Filler bits is roughly same as block size increase (spacing increases as block size increases).
Multiple parallelism values are available (block sizes are spaced an integer bytes apart).
Following 188 byte-aligned interleaver sizes spaced in a semi-log manner are selected with
approximately 3% filler bits:
K=
40 + 8t
if 0 t 59 (40512 in steps of 8 bits)
512 + 16t if 0 < t 32 (5281024 in steps of 16 bits)
1024 + 32t if 0 < t 32 (10562048 in steps of 32 bits)
2048 + 64t
if 0 < t 64 (21126144 in steps of 64 bits)
Maximum turbo interleaver size is increased from 5114 in UMTS to 6144 in LTE, such that a 1500
byte TCP/IP packet would be segmented into only two segments rather than three, minimizing
potential segmentation penalty and (marginally) increasing turbo interleaver gain.
1.4.6. Rate-Matching
Rate-Matching (RM) algorithm selects bits for transmission by puncturing and/or repetition, based on
the available physical resources. RM should send as many new bits as possible in retransmissions to
maximize Incremental Redundancy (IR) HARQ gains.RV = 0 starts at an offset relative to the
beginning of the CB(Code Block) to enable systematic bit puncturing on the first transmission.Circular
buffer RM was selected for LTE as it generates puncturing patterns simply and flexibly for any
arbitrary code rate, with excellent performance.
Fig 2.4.6 Encoder and Redundancy version generation

Each of the three output streams of the turbo coder (systematic part, parity0, and parity1) is
rearranged with its own interleaver (Sub-block interleaver). The 12 tail bits are distributed equally into
the three streams, resulting in sub-block size Ks = K + 4, where K is the QPP interleaver size. Then,
an output buffer is formed by concatenating the rearranged systematic bits with the interlacing of the
two rearranged parity streams. For any desired code rate, the coded bits for transmission are simply
read out serially from a certain starting point in the buffer, wrapping around to the beginning of the
buffer if the end of the buffer is reached.
A Redundancy Version (RV) specifies a starting point in the circular buffer to start reading out bits.
Different RVs are specified with different starting points to enable HARQ operation. RV = 0 is selected
Page 52
for initial transmission to send as many systematic bits as possible. The scheduler can choose
different RVs on transmissions of the same packet to support both IR and Chase combining HARQ.
Turbo code tail bits are uniformly distributed into the three streams, with all streams the same size.
Each sub-block interleaver is based on the traditional row-column interleaver with 32 columns, and a
simple length-32 intra-column permutation.
The bits of each stream are written row-by-row into a matrix with 32 columns (rows
determined by the stream size), with dummy bits padded to the front of each stream to
completely fill the matrix.
A length-32 column permutation is applied and the bits are read out column-by-column to
form the output of the sub-block interleaver, [0, 16, 8, 24, 4, 20, 12, 28, 2, 18, 10, 26, 6, 22, 14, 30, 1, 17, 9,
25, 5, 21, 13, 29, 3, 19, 11, 27, 7, 23, 15, 31]
This sub-block interleaver first puts all the even indices and then all the odd indices into the
rearranged sub-block.
A small percentage of systematic bits are punctured in an initial transmission to enhance performance
at high code rates. With the offset, RV = 0 results in partially systematic codes that are self-decodable
at high coding rates, avoiding the catastrophic puncturing patterns.
After interleaving, bits are read column-by-column starting from a column top RV location. This
enables efficient HARQ operation, because CB operation can be performed without requiring an
intermediate step of forming any actual physical buffer. For any combination of the 188 stream sizes
and 4 RV values, the desired codeword bits can be equivalently obtained directly from the output of
the turbo encoder using simple addressing based on sub-block permutation.
The buffer looks like Virtual Circular Buffer (VCB) and this allows Systematic Bit Puncturing (SBP) by
defining RV = 0 to skip the first two systematic columns of the CB, leading to approximately 6%
punctured systematic bits. Thus, with systematic bit puncturing and uniform spaced RVs, the four RVs
start at the top of columns 2, 26, 50 and 74.
1.4.7. HARQ in LTE

The physical layer in LTE supports HARQ on DL and UL shared channels, with separate control
channels to send the associated ACK/NACKs. In FDD, eight Stop-And-Wait (SAW) HARQ processes
are available in UL/DL with Round-Trip Time (RTT) of 8 ms, each HARQ identified with a unique three
bit HARQ process IDentifier (HARQ ID), and requires a separate soft buffer allocation in the receiver
for combining the retransmissions. There are several fields in DL control to aid HARQ operation:
New Data Indicator (NDI): toggled whenever a new packet transmission begins;
Redundancy Version (RV): indicates the RV selected for the transmission or retransmission;
MCS: modulation and coding scheme.
Fig 2.4.7 MAC HARQ sequence

Downlink HARQ is asynchronous and adaptive,and therefore every DL transmission is accompanied
by explicit signalling of control information. The UL HARQ is synchronous, and either non-adaptive or
Page 53
adaptive. The UL non-adaptive HARQ requires a predefined RV sequence 0, 2, 3, 1, 0, 2, 3, 1, . . . for

successive transmissions of a packet due to the absence of explicit control signalling. For adaptive
HARQ, RV is explicitly signalled. The redundancy version (or modulation) is combined with other
control information to minimize control signalling overhead.
Limited Buffer Rate Matching (LBRM)
The required UE HARQ soft buffer size is the total memory (over all HARQ processes) required for
LLR storage to support HARQ operation. Aim of Limited Buffer Rate Matching (LBRM) is to reduce
the required UE HARQ soft buffer sizes while maintaining peak data rates and system performance.
LBRM simply shortens the length of the VCB of code block segments for certain larger size of
Transport Block8 (TB), with RV spacing being compressed accordingly. The effective mother code
rate for a TB depends on TB size and allocated UE soft buffer size.
For example, for FDD & lowest categories of UE, LBRM does not result in any shortening of the soft
buffer. For higher UE categories, the soft buffer size is calculated assuming 8 HARQ processes and a
50% buffer reduction, which corresponds to a mother code rate of 2/3 for the largest TB. Since
eNodeB knows the soft buffer capability of the UE, it only transmits those code bits out of the VCB
that can be stored in the UEs HARQ soft buffer for all (re)transmissions of a given TB.
1.4.8. Coding for Control Channels in LTE

Unlike data, PDCCH and PBCH is coded with convolutional code, as code blocks are significantly
smaller and turbo coding complexity is not worthwhile. PDCCH is critical from a decoding complexity
point-of-view, since UE must decode a lot of potential control channel locations. Both channels carry a
relatively small number of bits, making tail bits a more significant overhead. Therefore, it is decided to
adopt a tailbiting convolutional code for LTE, but, to limit complexity, convolutional code with only 64
states is used.
Convolutional code offers slightly better performance for target block sizes. The initial value of the
shift register of the encoder is set to the values corresponding to the last six bits in the input stream so
that the initial and final states of the shift register are the same and decoder can utilize a Circular
Viterbi Algorithm, with decoding complexity approximately twice that of Viterbi decoder with two
iterations.
Rate-matching for convolutional code is similar circular buffer method as for turbo code. A 32-column
interleaver is used, with no interlacing in the circular buffer (three parity streams are concatenated in
the circular buffer). This structure gives good performance at higher code rates as well as lower code
rates, and no need for an additional (different) R = 1/2 generator polynomial. With small information
words, block codes lend themselves well to a maximum likelihood decoding approach, like for
PCFICH or PHICH.
1.4.9. General structure for downlink physical channels

For every downlink physical channel, there are following processing done after interleaving. The steps
are:
- scrambling of coded bits in each of the codewords to be transmitted on a physical channel
- modulation of scrambled bits to generate complex-valued modulation symbols
- mapping of the complex-valued modulation symbols onto one or several transmission layers
- precoding of the complex-valued modulation symbols on each layer for transmission on the
antenna ports
- mapping of complex-valued modulation symbols for each antenna port to resource elements
- generation of complex-valued time-domain OFDM signal for each antenna port
Scrambling
Modulation
mapper
Layer
mapper
Scrambling
antenna
ports
layers
codewords
Modulation
mapper
Resource
element mapper
OFDM signal
generation
Resource
element mapper
OFDM signal
generation
Precoding
Fig 2.4.9 - Overview of physical channel processing.

Page 54
1.4.10. Scrambling
For each codeword q of M bits shall be scrambled prior to modulation, resulting in a block of
~
scrambled M bits according to b ( q) (i) b (q) (i) c (q) (i) mod 2 , where the scrambling sequence is
c ( q ) (i) . The scrambling sequence generator shall be initialised at the start of each subframe, where
the initialisation value of cinit is
cell
n
214 q 213 ns 2 29 N ID
for PDSCH
cinit RNTI 9
MBSFN
ns 2 2 N ID
for PMCH
Up to two codewords can be transmitted in one subframe, i.e., q 0,1.
1.4.11. Modulation
For each codeword q of M scrambled bits, shall be modulated as using one of the modulation
schemes of QPSK, 16QAM or 64QAM, resulting in a block of complex-valued Msymb modulation
symbols.
1.4.12. Layer mapping

Modulated symbols for each codewords to be transmitted are mapped onto one or several layers.
Modulation symbols are mapped onto v number of layers with each layer having Msymb(v). There will
layer
be one symbol transmitted in each layer - x(i) x (0) (i) ... x ( 1) (i) , i 0,1,...,M symb
1 .
Layer mapping for transmission on a single antenna port

For transmission on a single antenna port, a single layer is used, 1 , and x (0) (i) d (0) (i) with
layer
(0)
.
M symb
M symb
Layer mapping for spatial multiplexing

Number of layers is less than or equal to the number of antenna ports P used for transmission.
Single codeword mapped to multiple layers is only applicable when the number of cell-specific RS is
four or when the number of UE-specific RS is two or larger.
Layer mapping for transmit diversity

For transmit diversity, there is only one codeword and the number of layers is equal to the number
of antenna ports P used for transmission of the physical channel.
1.4.13. Precoding
layer
Precoder takes as input a block of vectors x(i) for v layers ( i 0,1,...,M symb
1 ) from layer mapping and
(p)
ap
generates a block of vectors y (i) for p antenna ports each ( i 0,1,...,M symb
1 ) to be mapped onto
resources, where y ( p ) (i) represents the signal for antenna port p . For transmission on a single
ap
antenna port, precoding is defined by y ( p) (i) x (0) (i) for each symbol i 0,1,...,M symb
1 ,
ap
layer
.
M symb
M symb
Page 55
Precoding for spatial multiplexing using antenna ports with CRS

Precoding for spatial multiplexing using antenna ports with cell-specific reference signals is only used
in combination with layer mapping for spatial multiplexing as. Spatial multiplexing supports two or four
antenna ports and the set of antenna ports used is p 0,1 or p 0,1,2,3 , respectively.
Precoding without CDD

Without cyclic delay diversity (CDD), precoding for spatial multiplexing is defined by
y (0) (i )
x (0) (i )
W (i )
y ( P 1) (i )
x ( 1) (i )
ap
ap
layer
where the precoding matrix W (i) is of size P and i 0,1,...,M symb
.
M symb
1 , M symb
Values of W (i) shall be selected among the precoder elements in the codebook configured in the
eNodeB and the UE.
Precoding for large delay CDD

For large-delay CDD, precoding for spatial multiplexing is defined by
y (0) (i )
x (0) (i )
W (i ) D(i )U
y ( P 1) (i )
x ( 1) (i )
The diagonal size- matrix D(i) supporting CDD and the size- matrix U are given for
different numbers of layers . A different precoder is used every vectors, where denotes the
number of layers.
Codebook for precoding and CSI reporting
For transmission on two antenna ports, p 0,1 , and for the purpose of CSI reporting based on two
antenna ports p 0,1 or p 15,16 , the precoding matrix W (i) is selected.
Precoding for transmit diversity

Precoding for transmit diversity is only used in combination with layer mapping for transmit diversity.
The precoding operation for transmit diversity is defined for two and four antenna ports.
Precoding for spatial multiplexing using antenna ports with UE-specific RS

Spatial multiplexing using antenna ports with UE-specific RS supports up to eight antenna ports. For
y (7 ) (i ) x (0) (i )
(8 )

y (i ) x (1) (i )
transmission on antenna ports, the precoding operation is defined by

where
(6 ) ( 1)
(i ) x
(i )
y
ap
ap
layer
.
M symb
i 0,1,...,M symb
1 , M symb
1.4.14. Mapping to resource elements

ap
For each of the antenna ports, the symbols y ( p) (0),..., y ( p) (M symb
1) shall conform to the downlink
power allocation and be mapped in sequence starting with y ( p ) (0) to resource elements k, l , when:
-
they are in RB assigned for transmission, and
Page 56
they are not used for PBCH, PSS, SSS, MBSFN RS, UE-specific RS with PDSCH, and
they are not used for CRS, with the number of antenna ports for and the frequency shift of cellspecific RS unless other values for these parameters are provided, and
- they are not to be used for CSI RS, with zero power and non-zero power CSI RS, and the DCI
associated with DL uses the C-RNTI or semi-persistent C-RNTI, and
- they are not part of RB pair, carrying an EPDCCH associated with PDSCH, and
- l in the first slot in a subframe fulfils l lDataStart .
If the DCI uses C-RNTI or semi-persistent C-RNTI and transmit diversity is used, resource elements
in an OFDM symbol assumed by the UE to contain CSI-RS shall be used in the mapping above if and
only if all of the following criteria are fulfilled:
- there is an even number of RE for symbol in each RB assigned for transmission, and
- symbols y ( p ) (i) and y ( p ) (i 1) , where i is an even number, can be mapped to resource
elements k, l and k n, l in the same OFDM symbol with n 3 .
The mapping to RE k, l on antenna port p are in increasing order of first the index k over the
assigned RB and then the index l , starting with the first slot in a subframe.
Finally, the above mapping will be passed on to the stage of IFFT with CP signal insertion, then
concert parallel to serial and then transmit the signal out of the antenna.
Page 57
Page 58
1.5 MIMO Techniques

1.5.1. Introduction to MIMO
Invention of Multiple-Input Multiple-Output (MIMO) came in mid 90s. MIMO was adopted first time
from Release 7 version of HSDPA and LTE was designed with MIMO as a key component from the
start.
Single-Input Single-Output (SISO) exploit time or frequency domain pre-processing and decoding of
the transmitted and received data respectively. The use of multiple antenna at either eNodeB or UE
(on DL or UL) requires signal precoding and detection. Here is an example how signals are precoded
in every path and decoded back.
Fig 2.5.1.1 MIMO Precoding illustration

In multi-antenna enabled base station with a single antenna UE, uplink is SIMO and downlink is
MISO. When multi-antenna UE is used, it is called MIMO, also SIMO and MISO also is used within
MIMO definition. A point-to-point multiple-antenna link between eNB and one UE is referred as
Single-User MIMO (SU-MIMO), Multi-User MIMO (MU-MIMO) features several UEs communicating
simultaneously with a common eNB using the same frequency- and time-domain resources.
In a multicell context, neighbouring eNBs sharing their antennas in virtual MIMO to communicate with
the same set of UEs in different cells will be termed multicell multi-user MIMO.
There are basically three advantages of MIMO over SISO (a)Diversity gain, (b)Array gain and
(3)Spatial multiplexing gain.
Diversity gain corresponds to mitigation of multipath fading, by transmitting or receiving over multiple
antennas at which fading is sufficiently decorrelated. It is expressed in terms of an order of number of
effective independent diversity branches or to slope of BER curve (function of SNR or link budget
gain). Gain is related to improvement of statistics of instantaneous SNR, array gain and multiplexing
gain.
Array gain corresponds to a spatial version of matched-filter gain in time-domain receivers.
Multiplexing gain refers to the gain where data of multiple users is multiplexed and separated by
orthogonal spreading codes, timeslots or frequency assignments. MIMO multiplexing has no extra
cost of bandwidth expansion; but needs added antennas and signal processing complexity.
MIMO Signal Model

Page 59
MIMO signal Y is a matrix of size NxT precoded signals transmitted from N distinct antennas over T
symbol durations (or, in frequency-domain T subcarriers). Thus the n-th row of Y corresponds to nth
transmit antenna signal. Let H be MxN channel matrix modelling the propagation from each of the N
antennas to any one of the M receive antennas, on a arbitrary subcarrier, for T symbol durations.
Then the MxT signal received over T symbol durations, R = HY + N where N is the noise matrix of
dimension MxT over all M receiving antennas.
Fig 2.5.1.2 MxN Transmission Vector

The ith column of H (hi) can be receive spatial signature of ith transmitting antenna
The jth row
of H (hj) can be transmit spatial signature of jth receiving antenna.
The hij can be the signal transmitted by antenna i and received by j.
Mapping the symbols to the transmitted signal
X = (x0,x1,.xp) symbols are sent to M receivers over T symbol durations, mapped to the transmitted
signal Y before launching into the air. Mapping function XY(X) determines MIMO methods results,
giving specific diversity, array and multiplexing gains. Spatial rate of MIMO = P/T. The symbols X may
correspond to data of one (SU-MIMO) or multiple users (MU-MIMO).
1.5.2. Single-User (SU-) MIMO Techniques

Optimal Transmission over MIMO Systems
Optimal transmission requires channel-dependent precoder, taking the roles of both transmit
beamforming and power allocation across transmitted streams, and a matching receive beamforming
structure. Full channel knowledge is required at the transmit side. Let P=NxT symbols sent separated
into N streams (or layers) of T symbols each. Stream i xij= [xi,1 xi,2 xi,3 xi,T].
Each stream may adopt a distinct code rate and modulation. The transmitted signal can now be
written as:
Y(X) = VP.X Where
X = x1,1, x,1,2, .. .. .. ,x1,T,
x2,1, x,2,2 .. .. .. X2,T
------- ---xN,1, x,N,2 .. .. .. xN,T and
VP= Precoding Matrix = singular vector beamforming + waterfilling power allocation
V = NxN transmit beamforming matrix, and
P = NxN diagonal power allocation matrix with pi as its ith element, (pi = ith stream power
allocated).
2 H H
CMIMO = log2 (I + HVP V H ) = capacity of the MIMO channel in bps/Hz
Singular vector beamforming: It means that V should be a unitary matrix such that H = U_VH is
Singular-Value Decomposition (SVD) of the channel matrix H and ith right singular vector of H (ith
column of V) is used as a transmit beam-forming vector for ith stream.
At receiver, optimal beamformer for ith stream is the ith left singular vector of H, (ith row of UH):
uHi R = ipi [xi,1, xi,2, . . . , xi,T ] + u(H)i*N
where i is the ith singular value of H.
2
Waterfilling power allocation: It is the optimal power allocation pi = [ 1/( i) ]+
where [x]+ is equal to x if x is positive and zero otherwise.
Page 60
is water level, such that total power constraint is satisfied.
Fig 2.5.2.3 Waterfilling method

The optimal SU-MIMO multiplexing uses SVD-based transmit and receive beamforming to
decompose MIMO channel into a number of parallel non-interfering subchannels (eigen-channels),
each one with an SNR (singular value i based) and chosen power level pi .
The philosophy of optimal power allocation across the eigen-channels is not to equalize the SNRs,
but to render them more unequal, by pouring more power into the better eigen-channels, while
allocating little power (or even none at all) to the weaker ones because they are seen as not
contributing enough to the total capacity. In practice this is done by selecting a suitable Modulation
and Coding Scheme (MCS) for each stream.
Beamforming with Single Antenna Transmitter or Receiver
In a single antenna, multiplexing of more than one data stream is not possible.
In single-stream receive beamforming, N = 1 andM >1, one symbol is transmitted at a time (P = T =
1), and Y(X) = X = x, one symbol to be sent.
Received signal vector R = Hx + N. Receiver combines the signals from M antennas through weights
w =[w1, . . . , wM].
Received signal after combining z = wR = wHs + wN.
After receiving channel estimate CQI, it can set beamforming vector w to maximize the received
SNR, by Maximum Ratio Combining (MRC) w = HH, a spatial version of matched filter. Cancellation of
an interfering signal can also be achieved, by selecting beamforming vector to be orthogonal to the
channel from the interference source. Here is beamforming and interference cancellation concept:
Fig 2.5.2.4 Beamforming Illustration
Page 61
The MRC provides a factor of M improvement in received SNR compared to M = N = 1. Array gain
=10log10(M)dB in link budget.
In transmit beamforming, M =1 and N >1, Symbol-to-transmit-signal mapping has P=T=1, and Y(X)=
wx, where x is the symbol, w is transmit beamforming vector of size N1, computed based on channel
knowledge (based on feedback link). The SNR-maximizing is done by transmit MRC, seen as a
matched prefilter: w = HH/H where H enforces total power constraint across transmit antennas.
Transmit MRC pre-filter gain= 10log10(N)dB SINR improvement.
Spatial Multiplexing without Channel Knowledge at the Transmitter
WhenN >1 andM >1, multiplexing of MxN streams is possible. Assume M N, consider N streams,
each with different transmitter. If transmitter does not have knowledge of H, spatial multiplexing
scheme cannot be improved by channel-dependent precoder. Then precoder is simply the identity
matrix. Then symbol-to-transmit-signal mapping function P=NT and Y(X) = X
At receiver, linear and non-linear detection techniques are implemented to recover X. A lowcomplexity linear solution case, receiver superposes N beamformers w1, w2, . . . , wN. Detection of
stream [xi,1, xi,2, . . . , xi,T ] is achieved by applying wi = wiR = wiH.X + wiN.
Beamformer wi design can be interpreted as a compromise between single-stream beamforming and
cancelling of interference (created by other N 1 streams). Inter-stream interference is fully cancelled
by selecting the Zero-Forcing (ZF) receiver with W= f(w1, w2, ... wN).
For optimal performance, wi should strike a balance between alignment of hi and orthogonality of
signatures hk, by may be Minimum Mean-Squared Error (MMSE) receiver.
Beyond classical linear detection (ZF or MMSE) receivers, more advanced but nonlinear detectors
can be exploited at extra complexity like- Successive Interference Cancellation (SIC) detector and
Maximum LikelihoodDetector (MLD).
SIC treats individual channel-encoded streams, like layers which are peeled off one by one by a
processing sequence consisting of linear detection, decoding, remodulating, re-encoding and
subtraction from the total received signal R.
MLD selects the most likely set of all streams from R, by an exhaustive search procedure or a lowercomplexity equivalent such as sphere-decoding technique.
Multiplexing gain
Multiplexing gain multiplicative factor by which spectral efficiency is increased by a given scheme.
MIMO achieves gain for various antennas to experience a sufficiently different channel response, to
be sufficiently
decorrelated and linearly independent to allow for the channel matrix H to be invertible. There is a
limitation to MxN number of independent streams which may be multiplexed into the MIMO channel
(rank(H) streams). SU-MIMO between a four-antenna eNB and a dual antenna UE can, at best,
support multiplexing of two data streams, doubling UEs data rate compared with a single stream.
Diversity
A diversity-oriented design will feature some level of repetition between the entries of Y. For full
diversity, each symbol xi must be assigned to each transmit antennas at least once during T symbol
durations. Diversity symbol-to-transmit signal mapping function is called Space-Time Block Code
(STBC). In addition to STBC, orthogonality of matrix Y improves performance and easy decoding at
receiver, realized by Alamouti space-time code. The diversity order is equal to MxN. For this
transmission, no knowledge of channel and feedback is necessary.
Diversity versus multiplexing trade-off
There exists a compromise between reaching full beamforming gain in detection of a desired stream
of data and perfect cancelling of undesired, interfering streams. Similarly, there is a trade-off between
the number of multiplexed streams in MIMO channel and the amount of diversity each one of them
will enjoy.
In N streams over a N to M antenna channel, with M N, and using a linear detector, each stream will
enjoy a diversity order of M N + 1. Increasing the spatial load of MIMO (Rank streams) is akin to
increasing the user load in CDMA.
Page 62
1.5.3. Multi-User Techniques

Comparing Single-User and Multi-User MIMO
The set of MIMO techniques featuring data streams being communicated to (or from) antennas
located on distinct UEs through same time-frequency resources is called Multi-User MIMO (MUMIMO). The MU-MIMO scenario differs from single-user in many ways.
In MU-MIMO, K UEs are selected for simultaneous communication over the same time-frequency
resource, from a set of U active UEs in the cell. Typically, K <=U. Each UE is assumed to be
equipped with J antennas, so the selected UEs together form a set of M = KJ UE-side antennas.
Streams will be limited to min(M,N)for N antenna from eNB (if complete interference suppression is
intended), upper bound in MU-MIMO is typically dictated by number of base station antennas N.
MU-MIMO multiplexing benefits are preserved even in the case of low-cost UEs with a small number
of antennas. Thus eNB performs receive beamforming from several UEs on the UL and transmit
beamforming towards several UEs on DL. Another advantage in MU-MIMO is that the decorrelation
between the signatures of different UEs occurs naturally as the separation between such UEs is
typically large relative to the wavelength.
Techniques for Single-Antenna UEs
In MU-MIMO for single-antenna UEs, number of antennas available to a UE for transmission is
typically less than number available for reception.
In single antenna case, no precoding can be applied and each UE simply transmits an independent
message. Thus, if K UEs are selected for transmission in the same time-frequency resource, each UE
k transmitting symbol s(k) , the received signal at eNB, over T = 1 symbol period is R = H.X + N where
X =x(1) ...x(k). The columns of H correspond to the receive spatial signatures of different UEs. eNB
can recover transmitted symbol by applying beamforming filters, using MMSE or ZF solutions. Note,
no more than N UEs can be served (K N). MU-MIMO in UL is sometimes referred as Virtual MIMO,
UE has no knowledge of the simultaneous transmissions of the other UEs. On DL, eNB must resort to
transmit beamforming to separate data streams intended for the various UEs.
MU-MIMO in DL with single-antenna UEs: eNB transmits to K selected UEs simultaneously. Their
contributions are separated by multiple antenna precoding at eNB, based on channel knowledge.
Over T=1 symbol period, signal received by UEs 1 to K = R = r1 ...rK = HVP.X + N. The rows of H
correspond to the transmit spatial signatures of various UEs. V is the transmit beamforming matrix
and P is the (diagonal) power allocation matrix selected . To cancel out fully the inter-user
interference when K N, a transmit ZF beamforming solution is employed.
Techniques for Multiple-Antenna UEs
Idea of single-antenna UEs can be generalized to multiple antenna UEs. The multiple antennas may
simply be treated as multiple virtual UEs, one UE may receive or transmit more than one stream,
while at the same time spatially sharing the channel with other UEs. For instance, a four-antenna eNB
could communicate in a MU-MIMO fashion with two UEs equipped with two antennas each, two
streams per UE.
Additional UE antennas give extra degrees of freedom for better link between UE and eNB. Multiple
antennas at UE may be combined in MRC in DL, or space-time coding could be used in UL. Multiple
antennas will give better diversity.
Comparing Single-User and Multi-User capacity
Idealized channel model assumes that entries in channel matrix H are independently and identically
distributed Rayleigh fading. Lets perform channel measurements synchronously for two UEs moving
at vehicular speed in outdoor hilly environment with Line-Of-Sight (LOS) predominantly present.
The sum-rate capacity of a two-UE MU-MIMO system is compared with the capacity of an equivalent
MISO system serving a single UE at a time, employing beamforming. The base station has four
antennas and UE has a single antenna. Full CSI is assumed available in both cases.
Mean sum-rate of both schemes in both channels is taken, over all frames and all subcarriers and
subsequently normalized to bps/Hz. It can be seen that in both the ideal and the measured
channels,MU-MIMO yields a higher sumrate than SU-MISO in general. In fact, at high SNR, the
multiplexing gain of theMU-MIMO system is two while it is limited to one for the SU-MISO case.
However, for low SNR, the SU-MISO and MU-MIMO schemes perform very similarly and both are
worse than in the idealized i.i.d. channels. So, MU-MIMO should be used in better SINR environment.
Page 63
1.5.4. MIMO Schemes in LTE

The MIMO schemes adopted here for LTE relate to the DL unless otherwise mentioned. There are
few challenges which have to be overcome to earn the benefits of MIMO.
The full MIMO benefits (array gain, diversity gain and multiplexing gain) assume ideally
decorrelated antennas and full-rank MIMO channel matrices. In single-user case, antennas at
both eNB and UE are typically separated by between 1/2 wavelength to a few
wavelengths at most, which is very short in relation to eNB to UE. In a LOS situation, this will
cause a strong correlation between spatial signatures, limiting the use of multiplexing
schemes. But design itself may provide the necessary orthogonality properties even in LOS
situations. Two antennas (at both transmitter and receiver) that operate on orthogonal
polarizations (horizontal and vertical polarizations, +45 and 45 polarizations, which give a
twofold multiplexing capability even in LOS). Orthogonal polarizations at UEs may not always
be recommended as it results in non-omnidirectional beam patterns. Also, in SU-MIMO, the
condition of spatial signature independence can only be satisfied with rich random multipath
propagation.
Another source of discrepancy between theoretical MIMO gains and practically achieved
performance lies in the (in-)ability of the receiver to give right CSI frequently and perfectly,
but we know that they are limited and finite. This degrades the performance.In DL, MU-MIMO
relies on eNB to compute required transmit beamformer, which in turn requires CSI. If no
sufficient CSI available MU-MIMO gains disappear and SU-MIMO strategy becomes
optimal.SO, accurate Channel State Information (CSI) to be delivered by UE to eNB in a
resource-efficient manner is required. This requires use of appropriate codebooks for
quantization.
Another issue is the interaction between the physical layer and the scheduling protocol. In
both UL and DL cases, number of UEs served in MU-MIMO is limited to K = N, assuming
linear combining. Number of active users U will typically exceed K, right k users will have to
be scheduled for simultaneous transmission over a particular RB. This algorithm is not
specified in LTE and various approaches are possible. A combination of rate maximization
and QoS constraints may be considered. The choice of UEs that will maximize the sum-rate
is one that favours UEs exhibiting not only good instantaneous SNR but also spatial
separability among their signatures.
1.5.5. Single-User Schemes

Transmit Diversity Schemes
Transmit diversity is only defined for 2 and 4 transmit antennas, and one data stream, one codeword,
one transport block(TB). To maximize diversity gain, antennas need to be uncorrelated, well
separated relative to wavelength or have different polarization. Transmit diversity works better in low
SNR, low mobility, or low delay tolerant applications, or in signalling without feedback like MBMS,
PBCH, PSS and SSS. MIMO scheme is independently assigned for control channels and data
channels, and independently per UE in PDSCH.
Space-Frequency Block Codes (SFBCs)
If PHY channel is configured for transmit diversity with two eNodeB antennas, SFBC is used. SFBC is
a frequency-domain version of Space-Time Block Codes (STBCs or Alamouti codes). Codes are
designed so that the transmitted diversity streams are orthogonal and achieve the optimal SNR with a
linear receiver. Such orthogonal codes only exist for the case of two transmit antennas.
STBC always operates on pairs of adjacent symbols in time domain, not straightforward for LTE. For
SFBC, symbols transmitted from two eNodeB antenna ports on each pair of adjacent subcarriers are
defined as:
Y(0,1) y(0,2)
= x1 x2
Y(1,1) y(1,2)
-x2 x1
Where y(p,k) denotes the symbols transmitted from antenna port p on the kth subcarrier. Since no
orthogonal codes exist for antenna configurations beyond 2 2, SFBC has to be modified in order to
apply it to the case of 4 transmit antennas. In LTE, this is achieved by combining SFBC with FSTD.
Page 64
Frequency Switched Transmit Diversity (FSTD) and its Combination with SFBC
General FSTD schemes transmit symbols from each antenna on a different set of subcarriers. In
practice in LTE, FSTD is only used in combination with SFBC for the case of 4 transmit antennas, to
provide a suitable transmit diversity scheme where no orthogonal rate-1 block codes exists. Mapping
of symbols to antenna ports is different in 4 transmit-antenna case compared to the 2 transmitantenna SFBC scheme. This is because RS density on third and fourth antenna ports is half of first
and second antenna ports, and hence CSI accuracy may be lower on third and fourth antenna ports.
1.5.6. Beamforming Schemes

LTE differentiates between two transmission modes which may support beamforming for the PDSCH:
Closed-loop rank 1 precoding. Its a special case of SU-MIMO spatial multiplexing. UE feeds CSI
to eNodeB to indicate suitable precoding to apply for beamforming operation.
UE-specific RSs. UE does not feed back any precoding-related information. eNodeB tries to
deduce using Direction Of Arrival (DOA) estimations from UL, for which calibration of eNodeB RF
paths may be necessary.
This is mainly a mechanism to extend cell coverage by focussing eNodeB power in direction UE is
located.
It has following properties:
Implemented by an array of closely-spaced antenna elements for directional transmissions.
Signals from different antenna elements are phased appropriately so that they all add up
constructively at UE.
eNodeB ensurs that the beam is correctly directed.
A UE would not really be aware that it is receiving a directional transmission rather than a
cellwide transmission. To the UE, the phased array of antenna ports appears as just one
antenna.
Beamforming Issues
While using beamforming based on UE-specific RS, channel quality experienced by UE will
be different (better) than from CRS. eNodeB cannot rely on UE CQI feedback from UEspecific RS, when it is configured for UE-Specific RS. So, UE will derive the CQI using CRS.
eNB will be able to set an offset for CQI from UE-Specific RS to CRS, and then CRS then
may not be needed later.
It allows the possibility to use beamforming for UEs near cell edge, while other antenna ports
may be used for SU-MIMO spatial multiplexing to deliver high data rates to UEs closer to the
eNodeB.
Beamforming can only be applied to the PDSCH and not to PUCCH and controls. Hence,
overall cell range may still be limited by the range of control channels. PUCCH code rate
could be reduced when beamforming is applied to the PDSCH to extend cell range.
If the beamforming is intermittent, the strong intermittent interference may disturb CQI
reporting accuracy in adjacent cells.
1.5.7. What is Spatial Multiplexing?
Different streams (Layers) are generated by spatial multiplexing, which is a mapping of

symbols onto the transmit antenna ports. Each layer is identified by a (precoding) vector of
size equal to the number of transmit antenna ports and can be associated with a radiation
pattern.
The rank of the transmission is the number of layers transmitted.
A codeword is an independently encoded data block, a single Transport Block (TB) delivered
from MAC layer in transmitter to physical layer protected with CRC.
Number of codewords (max 2) <= number of layers<= number of antenna ports.
A single codeword may be mapped to all available layers, or multiple codewords are each mapped to
one or more different layers. There is separate HARQ ACK/NACK feedback for each codeword. In
LTE, at most two codewords are used, even if four layers are transmitted. Codeword to-layer mapping
is static. All RBs belonging to the same codeword use the same MCS, even if a codeword is mapped
to multiple layers.
Page 65
1.5.8. Precoding
PDSCH transmission modes for open-loop spatial multiplexing and closed-loop spatial multiplexing
use precoding from a defined codebook to form transmitted layers. Codesbooks are a set of
predefined precoding matrices, with size being a trade-off between signalling bits required to indicate
PMI and suitability of resulting transmitted beam direction.
Fig 2.5.8.1 PMI feedback and Precoding adjustment

In closed-loop spatial multiplexing, UE feeds back to eNodeB most desirable entry from a predefined
codebook. Preferred precoder matrix would maximize capacity. Some important properties of the LTE
codebooks are as follows:
Constant modulus property. Precoders in LTE are mostly pure phase corrections no amplitude
changes, with exception of identity matrix as the precoder. Identity precoder may completely switch
off one antenna on one layer, as long as the net effect across the layers is still constant modulus to
the PA.
Nested property. Codebooks of different ranks are arranged, so that lower rank codebook is subset
of the higher rank codebook vectors. It ensures that the precoded transmission for a lower rank is a
subset of the precoded transmission for a higher rank. For example, if a specific index in the
codebook corresponds to the columns 1, 2 and 3 in case of a rank 3 transmission, then the same
index in rank 2 transmission must consist of either columns 1 and 2 or columns 1 and 3.
Minimal complex multiplications. The 2-antenna codebook consists entirely of a QPSK alphabet,
all codebook multiplications use only 1 and j . The 4-antenna codebook does contain some QPSK
entries which require a 2 magnitude scaling as well. Here are precoding examples with diversity and
spatial multiplexing respectively.
Fig 2.5.8.2 Codebook example
Page 66
Fig 2.5.8.3 Codebook for 2 antennas
1.5.9. Cyclic Delay Diversity (CDD)

In open-loop spatial multiplexing, feedback from UE indicates only the rank of the channel, and not a
preferred PMI. In this mode, if rank for PDSCH >1, LTE uses Cyclic Delay Diversity (CDD). CDD
involves transmitting same set of symbols on same subcarriers from multiple transmit antennas, with
a different delay on each antenna. The delay is applied before CP is added, guaranteeing the delay is
cyclic over the FFT size.
Adding a time delay = applying a phase shift in frequency. Same time delay is applied to all
subcarriers, phase shift will increase linearly across the subcarriers. Each subcarrier will experience a
different beamforming pattern as non-delayed subcarrier interferes constructively or destructively with
the delayed version from another antenna. The diversity effect of CDD therefore arises from the fact
that different subcarriers will pick out different spatial paths in the propagation channel, thus
increasing the frequency-selectivity of the channel. It helps to ensure that any destructive fading is
constrained to individual subcarriers rather than affecting a whole TB. It is beneficial if CSI is
unreliable.
Delay is added before the CP means, any delay value is used without increasing overall delay spread.
The time-delay/phase-shift equivalence means that the CDD operation can be implemented as a
frequency-domain precoder. The implementation designer can choose whether to implement CDD in
the time domain or the frequency domain.
CDD is discussed so far for rank-1 transmission, single layer. In practice, CDD is only applied in LTE
when PDSCH rank>1. Each layer benefits independently from CDD as for a single layer. For
multilayer CDD operation, the mapping of the layers to antenna ports is carried out using precoding
matrices selected from the spatial multiplexing codebooks.
In the case of 2 transmit antenna ports, the predetermined spatial multiplexing precoding matrix W is
always the same (identity matrix). In the case of 4 transmit antenna ports, different precoding
matrices are used from the 4 transmit antenna port codebook where is the transmission rank.
Feedback Computation and Signalling
In addition to CQI, UE can be configured to report PMIs and RIs. If UE knows applicable PMI, and
channels estimation, it can determine and report most suitable W(PMI, precoder), which maximizes
aggregate data bits which could be received across all layers. eNodeB may restrict precoders
(codebook subset restriction) UE may evaluate and report. For each PDSCH to UE, eNodeB indicates
which precoder is used.
RBs in each is configurable by eNodeB.UE indicates RI for best DL data rate, and eNodeB indicates
rank being used for a PDSCH transmission.
Thus, CQI, PMI and RI are used in SU-MIMO in LTE downlink.
Page 67
1.5.10. Multi-User Schemes

Availability of accurate CSIT is required for MU-MIMO. Transmit antenna diversity or SU-MIMO
doesnt requires such a sophisticated CSI feedback.
Precoding Strategies and Supporting Signalling
MU-MIMO is particularly beneficial for increasing total cell throughput in the downlink:
When eNodeB has N Tx antennas and UN UEs in cell, full multiplexing gain N can be achieved even
when each UE has only single antenna, by Spatial Division Multiple Access (SDMA) based on linear
precoding for transmit beamforming. Moreover, U UEs in the cell is large, MU-MIMO throughput can
be gained by exploiting multi-user diversity through relatively simple UE-selection mechanisms.
Benefits depend on CSIT that the eNodeB receives from each UE.
In SU-MIMO, even few feedback bits per antenna is beneficial in steering transmission more
accurately towards UE. In MU-MIMO DL, CSI does affect the multiplexing gain, as it is essentially
interference-limited. Hence CSI is more important for MU-MIMO than for SU-MIMO. But too many
feedbacks will outweigh the system throughput. So a small fraction of UEs are selected for MUMIMO, rather than many.
B
One solution to too many feedback is to utilize a codebook of Nq = 2 N-dimensional vectors and
sending to eNB a B-bit index from codebook, and CQI for MCS selection.
Two main precoding techniques emerged as MU-MIMO candidates for LTE, both relying on a
codebook-based limited feedback concept.
a. Codebook of Unitary Precoding (UP) matrices (Technique 1)
Codebook contains L = Nq/N predefined unitary beamforming matrices of size N N. For each
beamforming matrix in codebook, each UE computes an SINR for each N beamforming vectors in
the matrix, assuming that the other N 1 vectors are used for interference to other UEs. Overall,
the UE computes Nq SINRs and signals back to eNodeB the codebook index corresponding to
the best SINR and the value of this SINR. eNodeB then selects beamforming matrix and schedule
UEs with suitable MCS for each UE to maximize cell throughput. In this scheme limited set of
unitary matrices for precoding are available, and multiplexing gain is maximum when enough UEs
are in the cell orthogonal channel signatures matching codebook matrices. This happens only for
a large density of UEs.
b. Codebook for Channel Vector Quantization (CVQ) (Technique 2)
B
In this case, codebook contains Nq = 2 unit-norm quantization vectors and is used by each UE to
quantize N-dimensional vector of channel measurements. Before quantization, the channel vector
is normalized by its amplitude, such that the quantization index captures information regarding
only the direction of the channel vector. The UE then feeds back this quantization index along
with a real number representing an estimate of its SINR, which depends on the amplitude of the
channel and the directional quantization error. In this case the UE does not know the set of
possible beamforming matrices in advance. The eNodeB utilizes this feedback information
collected from the UEs to select the UEs for transmission and to construct a suitable beamforming
matrix, for example according to a zero-forcing beamforming criterion. One simple yet effective
option is to combine UE-selection with naive zero-forcing precoding. In such a scheme the
eNodeB has the flexibility to design the precoding matrix making use of the CSIT provided by
UEs; however, a signalling mechanism has to be devised to send back to the UEs enough
information about the designed precoder to allow successful data demodulation. One effective
means for such signalling is the use of at least one precoded (or UE-specific) reference symbol
for each precoding vector being used.
In the limit of a large population, a zero-forcing beamforming solution converges to a unitary precoder
because, there will be N UEs with good channel conditions. On the other hand, issue with unitary
precoding is that for large codebooks it achieves a multiplexing gain max by one (i.e. same as TDM
between UEs). This is because, if p = 1/L = N/2B is the probability that a UE selects a given
beamforming matrix in the codebook, then the probability of l out of U UEs selecting the same matrix
is (p, U) and mean value l = Up. Hence the average number of UEs selecting the same beamforming
matrix decreases exponentially with the codebook size in bits. Eventually, for large B, if U is kept
constant, only a single UE will be allocated per subframe.
In the standard, included are: (1) format and procedure for feedback from UEs, which may or may not
include method of calculating quantities to feed back; (2) Codebooks for feedback calculation and
Page 68
precoder selection by eNodeB; (3) Method and procedure for signalling the precoding weights to the
relevant UEs.
Calculation of Precoding Vector Indicator (PVI) and CQI
MU-MIMO supports only one codeword, to each UEs, we refer to precoding feedback as a Precoding
Vector Indicator (PVI) is used for both SU-MIMO and MU-MIMO.
Precoder and feedback calculation methods
Consider a codebook for feedback from the UEs, C = {C(0), . . . , C(L1)}, of L unitary matrices of size
NN, N = #transmit antennas & codebook size in number of vectors is Nq = L*N. Consider two
methods of calculating PVI from codebook and CQI.
1. In MU-MIMO precoding technique 1 (codebook of UP matrices), UE does following:
a. The SINR is computed for each vector in the codebook taken as UEs own precoding
vector. The interfering precoding vectors are assumed to be the other N 1 vectors in the
codebook.
b. Largest such SINR is used to report the CQI value, while useful precoding vector
represents PMI.
2. In MU-MIMO precoding technique 2 (codebook for CVQ), UE does following:
a. The measured channel vector is quantized to the nearest vector in the codebook in terms
of chordal distance.
b. SINR is estimated with assumptions:
o There are N1 interfering precoding vectors to the PVI vector (UE assumes that
zero-forcing beamforming will be used).
o The useful precoding vector is only slightly offset from the PVI vector.
Equivalence between channel vector quantization and PVI calculation
PVI and CQI is used by the eNodeB to construct PMI. In MU-MIMO precoding technique 1, a
precoder matched to the selected PVIs have to be orthogonal. In MU-MIMO precoding technique 2,
the PVIs of the selected UEs need not be exactly orthogonal and the eNodeB would be allowed to
adjust the precoder to minimize interference among UEs.
PDSCH transmission in MU-MIMO mode is based on selection of PMI from the same codebooks as
SU-MIMO. LTE 2-antenna codebook, and first eight entries in 4-antenna codebook, are DFT-based
codebooks. The main reasons for this choice are as follows:
Simplicity in the PVI/CQI calculation.
A DFT matrix nicely captures the characteristics of a MISO channel. If the first N rows are
taken from a DFT matrix of size Nq , each of the Nq column vectors that is obtained contains
the phases of a line-of-sight propagation channel from a uniform linear array with N elements
to a point in space located at a given angle with respect to the antenna array boresight.
The vectors in such a codebook can be grouped into L = Nq/N unitary matrices.
User Selection Mechanism
One important aspect of MU-MIMO is the selection of UEs, as throughput gain depends on exploiting
multi-user diversity.
Group the UEs that report orthogonal PVIs and select the group providing the highest total
throughput. Capacity formula R=k log(1+SINRk), which sums up the SINR(CQI) over the UEs in the
group.
Another greedy strategy consists of adding one UE at a time, as long as the additional UE increases
the overall throughput, for an achievable sum-rate. Let R(S) denote the achievable sum-rate when
UEs set S is selected.
Initialize S = { }, R(S) = 0.
(sum rate=0)
while |S| N do
(max user=N)
{ k = argmax R(S {k}) (k is not an element of S)
if R(S {k}) > R(S) update S = S {k},
else exit.
}
Receiver Spatial Equalizers
Commonly, three possible spatial equalizersat UE side may be used.
1. MMSE receiver.
Page 69
2.
3.
a. Receiver requires user to estimates N vectors of size J by estimating channel matrix

from CRS and PMI used. If precoded RS are provided, then directly estimated from
them.
Minimum Quantization Error (MQE) receiver.
a. Calculation does not require knowledge of PMI. It minimizes the vector quantization
error independently of the transmit beamformer, and maximizes the CQI.
Singular Value Decomposition (SVD) of the channel.
1. This receive beamformer depends only on the channel matrix Hk and is independent
of any CSI feedback.
1.5.11. Physical-Layer MIMO Performance

Consider two MU-MIMO unitary precoding and zero-forcing precoding with equal power distribution
across the selected UEs. Consider N = 4 antennas at eNodeB and M = 1,2 receive antennas per UE.
As codebook size increases, precoding shows degradation due to scheduling problem, too few UEs
select the same precoding matrix, making full-rank transmission less likely. Hence, multiplexing gain
for large codebooks falls to one.
For a large UEs, unitary precoding and zero-forcing will perform the same, as a set of N UEs can be
found with almost orthogonal channel signatures. The cell throughput with Zero-Forcing Equal Power
(ZFEP) slightly decreases with the granularity of the channel quantization for a number of reported
bits per index larger than 6. In fact, for a small codebook size UEs with very similar channel
signatures fall into the same quantization bin, and hence two such UEs are never selected
simultaneously.
Conversely, for larger codebooks two UEs with similar separation of their channel signatures may fall
into separate bins, and they may be selected for transmission as a result of the proportional fair
scheduling mechanism, thus causing slight degradation in throughput.
The performance degradation of zero-forcing beamforming may occur if the reported channel
feedback information is very inaccurate. Zero-forcing beamforming can be less robust against
inaccuracies in the channel representation compared to unitary precoding. This would be the case, if
the codebook size for the channel feedback is small or channels are spatially highly uncorrelated.
When it is not possible to provide sufficiently accurate channel information, the unitary precoding can
prove more robust and than a zero-forcing approach.
Page 70
2. Uplink Physical Layer

2.1. SC-FDMA Principles
To minimize intracell interference, orthogonal transmission by different UEs is required. It needs to
support adaptive data rates, frequency diversity, frequency-selective scheduling and spatial diversity
with MIMO. Since the UE size is small, it should have Low Peak-to-Average Power Ratio (PAPR) or
Cubic Metric (CM) for efficient Amplifier. To fulfil this criteria, Single-Carrier Frequency Division
Multiple Access (SC-FDMA) is chosen for UL.
SC-FDMA advantage over UMTS is that it achieves intra-cell orthogonality even in frequencyselective channels. However, a UE with code-multiplexed uplink also suffers increased CM/PAPR.
OFDMA could have been suitable for UL as well, except for low CM/PAPR. SC-FDMA combines
OFDM with the low CM/PAPR of single-carrier transmission schemes.
Like OFDM, SC-FDMA divides the bandwidth into multiple parallel subcarriers, with the orthogonality
between the subcarriers by Cyclic Prefix (CP) or guard period to prevent Inter-Symbol Interference
(ISI) between SC-FDMA information blocks. It transforms the linear convolution of the multipath
channel into a circular convolution to equalize the channel.
Fig 3.1 DFT Process

However, unlike OFDM, in SC-FDMA the signal modulated onto a given subcarrier is a linear
combination of all the data symbols transmitted at the same time instant. In each symbol period, all
the transmitted subcarriers of an SC-FDMA signal carry a component of each modulated data symbol.
This gives SC-FDMA its crucial single-carrier property, which results in low CM/PAPR than OFDM.
2.1.1. SC-FDMA Signal Generation (DFT-S-OFDM)

Generation of an SC-FDMA signal in the frequency domain uses a Discrete Fourier TransformSpread OFDM (DFT-S-OFDM) structure. The first step of DFT-S-OFDM SC-FDMA signal generation
is to perform an M-point DFT operation on each block of M QAM data symbols. Zeros are then
inserted among the outputs of the DFT to match the DFT size to an N-subcarrier OFDM modulator
(typically an Inverse Fast Fourier Transform (IFFT)).
Page 71
Fig 3.1.1.1 SC-FDMA Processing in Uplink

The zero-padded DFT output is mapped to N subcarriers, with the positions of the zeros determining
to which subcarriers the DFT-precoded data is mapped.
Usually N is larger than the maximum number of occupied subcarriers, thus providing for efficient over
sampling and sinc (sin(x)/x) pulse-shaping. The equivalence of DFTS-OFDM and a time-domaingenerated SC-FDMA transmission can readily be seen by considering the case of M = N, where the
DFT operation cancels the IFFT of the OFDM modulator resulting in the data symbols being
transmitted serially in the time domain. However, this simplistic construction would not provide any
oversampling or pulse-shape filtering.
Localized transmission.
The subcarrier mapping allocates a group of M adjacent subcarriers to a user. M <N results in
zero being appended to the output of the DFT spreader resulting in an up-sampled/interpolated
version of the original M QAM data symbols at the IFFT output of the OFDM modulator. The
transmitted signal is thus similar to a narrowband single carrier with a CP (equivalent to time-domain
generation with repetition factor L = 1) and sinc pulse-shaping filtering (circular filtering).
Distributed transmission.
The subcarrier mapping allocates M equally-spaced subcarriers (e.g. every Lth subcarrier). (L
1) zeros are inserted between the M DFT outputs, and additional zeros are appended to either side
of the DFT output prior to the IFFT (ML< N). As with the localized case, the zeros appended on either
side of the DFT output provide upsampling or sinc interpolation, while the zeros inserted between the
DFT outputs produce waveform repetition in the time domain. This results in a transmitted signal
similar to time-domain IFDMA with repetition factor L and sinc pulse-shaping filtering.
2.1.2. SC-FDMA Design parameters in LTE

SC-FDMA parameters chosen for the LTE uplink have been optimized under frequency-domain DFTS-OFDM signal generation.
Subframe duration
Slot duration
Subcarrier spacing
SC-FDMA symbol duration
CP duration 4.69 s all other symbols
Number of symbols per slot
Number of subcarriers per RB
1 ms
0.5 ms
15 kHz
66.67 s
Normal CP: 5.2 s first symbol in each slot,
Extended CP: 16.67 s all symbols
7 (Normal CP)
6 (Extended CP)
12
An important feature of the LTE SC-FDMA parameterization is that the numbers of subcarriers which
can be allocated to a UE for transmission are restricted such that the DFT size in LTE can be
constructed from multiples of 2, 3 and/or 5. This enables efficient, low complexity mixed-radix FFT
implementations.
The same basic transmission resource structure is used for the uplink as for the downlink: a 10 ms
radio frame is divided into ten 1 ms subframes each consisting of two 0.5 ms slots. It uses the same
Page 72
15 kHz subcarrier spacing. Smallest unit of resource being a Resource Element (RE), consisting of
one SC-FDMA data block length on one subcarrier. A Resource Block (RB) comprises 12 REs in the
frequency domain for a duration of 1 slot. A normal CP of duration 4.69 s and an extended CP of
16.67 s may be configured. Extended CP is beneficial for deployments with large channel delayspread characteristics, and for large cells.
The 1 ms subframe allows a 1 ms scheduling interval (or Transmission Time Interval (TTI)). However,
one difference from the downlink is that the uplink coverage is more likely to be limited by the
maximum transmission power of the UE. In some situations, when some Voice-over-IP (VoIP) packet
cannot be transmitted in a 1 ms subframe and segmentation may be required leading to more
signalling. A more efficient technique for improving uplink VoIP coverage at the cell edge is to use socalled TTI bundling, where a single transport block from the MAC layer is transmitted repeatedly in
multiple consecutive subframes, with only one set of signalling messages for the whole transmission.
The LTE uplink allows groups of 4 TTIs to be bundledin this way, in addition to the normal 1 ms TTI.
Uplink SC-FDMA parametrization for selected carrier bandwidths.
Carrier bandwidth (MHz)
1.4 3
5
10
15
20
FFT size
128 256 512 1024 1536 2048
Sampling rate: M/N 3.84 MHz
1/2 1/1 2/1 4/1 6/1 8/1
Number of subcarriers
72
180 300 600 900 1200
Number of RBs
6
15
25
50
75
100
Bandwidth efficiency (%)
77.1 90 90
90
90
90
In practice in LTE, all the uplink data transmissions are localized, using contiguous blocks of
subcarriers. Frequency-diversity is achieved by frequency hopping, which can occur both within one
subframe (at the boundary between the two slots) and between subframes. In the case of frequency
hopping within a subframe, the channel coding spans the two transmission frequencies, and therefore
the frequency diversity gain is maximized through the channel decoding process. The only instance of
distributed transmission in the LTE uplink (using an IFDMA-like structure) is for the Sounding
Reference Signals (SRSs) which are transmitted to enable the eNodeB to perform uplink frequencyselective scheduling.
System bandwidth is scalable from approximately 1.4 MHz up to 20 MHz with the same subcarrier
spacing and symbol duration for all bandwidths. Sampling rates resulting from FFT sizes are designed
to be small rational multiples of UMTS 3.84 MHz chip rate, for ease of implementation in a multimode
UE.
d.c. Subcarrier in SC-FDMA

No unused d.c. subcarrier is possible for SC-FDMA as it can affect the low CM/PAPR property of the
transmit signal. Direct conversion transmitters and receivers can introduce distortion at the carrier
frequency. The subcarriers are frequency-shifted by half a subcarrier spacing (7.5 kHz), resulting in
an offset of 7.5 kHz for subcarriers relative to d.c. Thus two subcarriers straddle the d.c. location. This
is the option used in uplink SC-FDMA.
Pulse Shaping
In SC-FDMA, there is no need for explicit pulse-shaping thanks to the implicit sinc pulse-shaping.
Nevertheless, an additional explicit pulse-shaping filter can further reduce the CM/PAPR, but at the
expense of spectral efficiency. As a result of this trade-off, additional pulse shaping is not specified in
LTE.
The important properties of the SC-FDMA transmission scheme used for the LTE uplink are derived
from its multicarrier OFDM-like structure with single-carrier characteristic. The multicarrier-based
structure gives the LTE uplink the same robustness against ISI as the LTE downlink, with lowcomplexity frequency-domain equalization being facilitated by the CP. At the same time, the DFTbased pre-coding ensures that the LTE uplink possesses the low CM required for efficient UE design.
Crucially, LTE uplink is designed to be orthogonal in the frequency domain between different UEs,
thus virtually eliminating the intra-cell interference associated with CDMA.
The parameters of the LTE uplink are designed to ensure maximum commonality with the downlink,
and to facilitate frequency-domain DFT-S-OFDM signal generation. The localized resource allocation
scheme of the LTE uplink allows both frequency selective scheduling and the exploitation of
frequency diversity, the latter being achieved by means of frequency hopping.
Page 73
Page 74
2.2. UL Physical Channel Structure

SC-FDMA provides separate physical channels for data and control. Channel structure is designed
efficiently use available frequency-time domain resources, and multiplexing data and control
signalling.MIMO introduces, closed-loop antenna selection, Spatial Division Multiple Access (SDMA)
or Multi-User Multiple-Input Multiple-Output (MU-MIMO). Uplink comprises of three physical channels
and two signals:
PUSCH Physical Uplink Shared CHannel
PUCCH Physical Uplink Control CHannel
PRACH Physical Random Access CHannel
DM RS DeModulation Reference Signal
SRS Sounding Reference Signal
Fig 3.2.0 UL Physical Channel mapping
2.2.1 Uplink Shared Data Channel Structure

The Physical Uplink Shared CHannel (PUSCH), which carries data from UL-SCH transport channel,
uses DFT-S-OFDM, information bits are Processed as follows:
1. Compute and append CRC, then
2. Coded with a turbo code of mother code rate r=1/3.
3. Rate is adapted to a suitable final code rate by rate-matching process,
4. Interleavingadjacent data symbols mapped first to adjacent symbols in time, then across
subcarriers.
5. Scrambled by a length-31 Gold code prior to modulation mapping,
6. OFDM modulation.
7. DFT-spreading,
8. Layer mapping and then Subcarrier (Resource element) mapping for each layer, and
9. Signal is frequency-shifted by half a subcarrier prior to transmission, to avoid distortion by d.c.
subcarrier being concentrated in one RB.
Page 75
Fig 3.2.1 PUSCH channel processing stages

The modulations supported are QPSK, 16QAM and 64QAM.
The baseband SC-FDMA transmit signal for SC-FDMA symbol is of the form,
for 0 t <(NCP, + N)Ts, where NCP= #samples of CP in SC-FDMA symbol (see Section 15.3), N =
RB
2048 is IFFT size, df = 15 kHz subcarrier spacing, Ts = 1/(Ndf ) is the sampling interval, N is
()
system BW in RBs, Nsc = 12 subcarriers per RB, k = k + [NRBNsc /2] and ak,l is the content of
subcarrier k on symbol l. For PUSCH data, ak,l (k=0,1,2.M-1) is obtained by DFT-spreading the
data QAM symbols, [d0,l, d1,l, . . . , dM1,l] to be transmitted on data symbol l,
A Hybrid Automatic Repeat reQuest (HARQ) is used synchronously, using N-channel stop and wait
(retransmissions occur in specific periodically-occurring subframes (HARQ channels)).
2.2.2 Scheduling in LTE SC-FDMA Uplink

In UL, both frequency-selective & non-frequency-selective scheduling are supported. Frequencyselective is based on eNodeB exploiting available channel knowledge to schedule a UE to transmit
using specific Resource Blocks (RBs) in frequency domain where the channel response is good. NonFrequency-Selective scheduling does not use frequency-specific channel knowledge, but rather aims
to benefit from frequency diversity during the transmission of each transport block.
Frequency-Selective Scheduling
Localized allocation is used in both slots of a subframe no frequency hopping during a subframe. RB
allocation and MCS are chosen based on location and quality in UL channel response. Timely
Page 76
channel quality information is needed at the eNodeB. SRS may be used to get quality info, whose
quality depends on SRS BW.
Frequency-Diverse or Non-Selective Scheduling
Frequency hopping of a localized transmission is used to provide frequency diversity (1)hopping
only between subframes (intersubframe hopping), or (2) hopping both between and within subframes
(inter- and intra-subframe hopping). hopping modes are configured by Cell-specific broadcast.
1. Intra-subframe hopping - frequency hop occurs at the slot boundary in the middle of a
subframe; this provides frequency diversity within a codeword (within TB).
2. Inter-subframe hopping - frequency diversity between HARQ retransmissions of TB, as the
frequency allocation hops every allocated subframe.
Either (1) a pre-determined pseudo-random frequency hopping pattern, or (2) an explicit hopping
offset signalled in the UL resource grant on the PDCCH.
For NRB < 50RBs, size of hopping offset is approximately NRB/2, &
For NRB >50 RBs, possible hopping offsets are NRB /2, and +- NRB/4..
In UL, resource grant indicated frequency hopping. Semiselective scheduling is when the frequency
resource is assigned for the first slot and frequency diversity is also achieved by hopping to a different
frequency in the second slot.
2.2.3 Uplink Control Channel Design

UL control signalling can be divided into two categories:
1. Data-associated. Always transmitted together with data & used in the processing of that
data. Examples: MCS, new data indicators (NDI), and MIMO parameters.
2. Data non-associated. Not associated with UL data, transmitted independently of UL data.
Examples: Scheduling Request (SR) for UL, HARQ (ACK/NACK) for DL, Channel Quality
Indicators (CQIs) and MIMO feedback (RI or PMI) for DL.
Low signalling latency achieved by short 1ms subframe, together with orthogonal nature of UL
multiple access necessitates centralized resource allocation by eNB. Consequently UL dataassociated control signalling is not necessary, as relevant information is already known to eNB.
Therefore only data non-associated control signalling is supported in UL.
When simultaneous PUSCH data and control signalling is scheduled for a UE, control signalling is
multiplexed with data prior to DFT spreading. PUCCH, is used by a UE to transmit any necessary
control signalling only in subframes in which the UE has not been allocated any RBs for PUSCH.
2.2.4 Physical Uplink Control Channel (PUCCH)

PUCCH is transmitted in edge frequencies of system bandwidth. To minimize resources needed for
PUCCH, we exploit frequency diversity: each PUCCH Tx in one subframe is comprised of a single
(0.5 ms) RB at or near one edge of BW, followed (in the second slot of the subframe) by a second RB
at or near the opposite edge of BW; together, the two RBs are referred to as a PUCCH region. This
design achieves a frequency diversity benefit of approximately 2 dB compared to transmission in
same RB in the subframe.
Narrow bandwidth of PUCCH in each slot (single RB) maximizes power per subcarrier for a given total
Tx power, and therefore helps to fulfil stringent coverage requirements. Link budget of a two-slot
narrowband transmission exceeds that of a one-slot wider-band Tx. Positioning the control regions at
the edges of the system bandwidth has a number of advantages:
Page 77
1. Frequency diversity achieved through frequency hopping is maximized by allowing hopping

from one edge of the band to the other.
2. Out-Of-Band (OOB) emissions are smaller if a UE is only transmitting on a single RB per slot
compared to multiple RBs. PUCCH regions can serve as a guard band between widerbandwidth PUSCH transmissions of adjacent carriers, and therefore can improve
coexistence.
3. It maximizes the PUSCH data rate, as the entire central band can be allocated to a single UE,
compared to if PUCCH was in the middle.
4. Imposes fewer constraints on the UL scheduling, both with and without inter-/intra-subframe
frequency hopping.
RBs (in each slot) that can be used for PUCCH transmission within cell is NRB(PUCCH) (puschHoppingOffset), indicated by SIB. PUCCH RBs per slot = number of PUCCH regions per subframe.
In case of odd numbers of PUCCH RB pair used, the unused RB pair may be used for PUSCH or
PUCCH may be hopped in that slot.
2.2.5 Multiplexing of UEs within a PUCCH Region

Control signalling from multiple UEs can be multiplexed into a single PUCCH region using Orthogonal
Code Division Multiplexing (CDM). One orthogonality technique between UEs is to use cyclic time
shifts of a sequence with suitable properties. In a PUCCH symbol, different cyclic time shifts of a
waveform are modulated with a UE-specific QAM symbol carrying the necessary control signalling
information, with 12 cyclic time shifts determining the number of UEs which can be multiplexed per
SC-FDMA symbol.
For just 1 or 2 bits (ACK/NACK) control transmissions, orthogonality is achieved between UEs by a
combination of cyclic time shifts within an SC-FDMA symbol and SCFDMA symbol time-domain
spreading with orthogonal spreading codes. CDM of multiple UEs is used rather than Time Domain
Multiplexing (TDM) because CDM enables the time duration of the transmission to be longer, which
increases transmitted energy per signalling message in case of a power-limited UE.
Thus, PUCCH uses frequency-domain code multiplexing (different cyclic time shifts) and/or timedomain code multiplexing (different orthogonal block spreading codes), providing an efficient,
orthogonal control channel which supports small payloads (up to 22 coded bits) from multiple UEs
simultaneously.
Page 78
2.2.6 Control Signalling Information Carried on PUCCH

PUCCH may carry:
1. Scheduling Requests (SRs).
2. HARQ ACK/NACK for PDSCH. One ACK/NACK bit for each codeword.
3. CQI, RIs and PMI. 20 bits are used per subframe for the CQI/PMI/RI.
Amount of control a UE can transmit in a subframe depends on symbols available for signalling data
excluding Ref Signal. Different PUCCH formats may be used depending on the information to be
signalled.
Physical mapping of the PUCCH formats to PUCCH regions is different. PUCCH CQI formats 2/2a/2b
are mapped on the band-edge RBs (PUCCH region m=0,1) followed by a mixed PUCCH RB (region
m = 2) of CQI format 2/2a/2b and format 1/1a/1b, and then by PUCCH format 1/1a/1b (eregion
m=4,5). PUCCH RBs available for CQI format 2/2a/2b, N2 RB, is broadcast in SI.
Table Supported uplink control information formats on PUCCH.
PUCCH Format
Format 1
Format 1a
Format 1b
Format 2
Format 2
Format 2a
Format 2b
Uplink Control Information (UCI)

Scheduling request (SR) (unmodulated waveform)
1-bit HARQ ACK/NACK with/without SR
2-bit HARQ ACK/NACK with/without SR
CQI (20 coded bits)
CQI and 1- or 2-bit HARQ ACK/NACK (20 bits) for extended CP only
CQI and 1-bit HARQ ACK/NACK (20 + 1 coded bits)
CQI and 2-bit HARQ ACK/NACK (20 + 2 coded bits)
2.2.7 CQI Transmission on PUCCH (Format 2)

The number of RS symbols per slot is a trade-off between channel estimation accuracy and code
rate. For small control bits with a low SNR, using more RS symbols (3 for format 1) is more beneficial
than being able to use a lower channel code rate. With larger control bits the required SNR increases,
and higher code rate (because RS symbols overhead is more) becomes more critical and favour
fewer RS symbols. Thats why PUCCH Format 2 uses symbols 1 and 5 for DMRS for normal CP
(symbol 3 only in extended CP).
10 CQI bits are channel coded with a rate 1/2 punctured(20,k)ReedMuller code gives 20 coded bits,
which are then scrambled (with length-31 Gold sequence) prior to QPSK modulation. One symbol is
transmitted on each of the 10 SC-FDMA symbols in the subframe by modulating a cyclic time shift of
the base RS sequence of length-12 prior to OFDM modulation. The 12 equally-spaced cyclic time
shifts allow 12 different UEs to be orthogonally multiplexed on the same CQI PUCCH RB.
The DMRS signal sequence (on the 2nd and 6th SC-FDMA symbols for the normal CP, or the 4th
symbol for the extended CP) is similar to the frequency domain CQI signal sequence but without the
CQI data modulation. In order to provide inter-cell interference randomization, cell-specific symbollevel cyclic time shift hopping is used, as described in Section 16.4. For example, the PUCCH cyclic
time shift index on SC-FDMA symbol l in even slots ns is obtained by adding (modulo-12) a pseudorandom cell-specific PUCCH cyclic shift offset to the assigned cyclic time shift nPUCCH RS. Intra-cell
interference randomization is achieved by cyclic time shift remapping in the second slot.
A UE is semi-statically configured by higher layer signalling to report periodically different CQI, PMI,
and RI types on CQI PUCCH using a PUCCH resource index n(2) PUCCH, which indicates both the
PUCCH region and the cyclic time shift to be used.
2.2.8 Multiplexing CQI and ACK/NACK on PUCCH

Simultaneous transmission of ACK/NACK and CQI by UE is enabled by UE-specific signalling. If
simultaneous is not enabled & UE needs to transmit ACK/NACK on the PUCCH where CQI report is
configured, CQI is dropped and only ACK/NACK is transmitted.
In subframes where eNB scheduler allows simultaneous CQI and ACK/NACK from UE, CQI and
ACK/NACK needs to be multiplexed in the same PUCCH RB. Here are various methods described.
Multiplexing of CQI and HARQ ACK/NACK Normal CP (Format 2a/2b)
To transmit ACK/NACK together with CQI (Format 2a/2b), ACK/NACK bits (which are not scrambled)
are modulated into a single symbol, dHARQ. A ACK is encoded as a binary 1 and NACK as 0.
Page 79
The dHARQ, is then used to modulate the second RS symbol (5th symb) in each CQI slot i.e.
ACK/NACK is signalled using RS. NACK is mapped to +1, resulting in a default NACK in case of DTX.
If UE fails to detect DL grant on PDCCH or it is DTX, it is as good as NACK and triggers DL
retransmission.
Multiplexing of CQI and HARQ ACK/NACK Extended CP (Format 2)
In extended CP (One RS symbol per slot), ACK/NACK is jointly encoded with CQI resulting in a (20,
kCQI + kACK/NACK) Reed Muller based block code. A 20-bit codeword is transmitted on PUCCH
using CQI channel structure. The largest number of information bits supported by the block code is
13, corresponding to kCQI = 11 CQI bits and kACK/NACK = 2 bits (1 for each codeword).
HARQ ACK/NACK Transmission on PUCCH (Format 1a/1b)

The PUCCH channel sends ACK/NACK without CQI. Three (two in case of extended CP) symbols are
used in the middle of the slot for RS, and rest for ACK/NACK. Due to small number of ACK/NACK
bits, three RS symbols are used to improve the channel estimation accuracy. Overall Physical layer
processing is shown here.
Fig 3.2.8 UE Transmitter and eNB Receiver stages

Both 1 and 2-bit acknowledgements are supported using BPSK and QPSK respectively. The HARQ
ACK/NACK bits (which are not scrambled) are BPSK/QPSK modulated according to the modulation
mapping resulting in a single ACK/NACK modulation symbol. ACK is encoded as 1 and NACK as 0.
The modulation mapping is the same formats 2a/2b.
Just as CQI transmission, the ACK/NACK symbol (which is phase-rotated by 90 degrees in the
second slot) is transmitted on each SC-FDMA data symbol by modulating a cyclic time shift of the
base RS sequence of length-12 prior to OFDM modulation. Time-domain spreading with orthogonal
spreading codes is used to code-division-multiplex UEs. Thus, a large number of UEs (data and RSs)
can be multiplexed on the same PUCCH RB using frequency-domain and time-domain code
multiplexing. For cyclic time shift multiplexing, cell-specific dPUCCH shift {1, 2, 3}, indicating 12, 6, or 4
shifts respectively, and selected based on expected delay spread in the cell.
For time-domain spreading CDM, spreading codes are limited by number of RS symbols, as
multiplexing capacity of RS is smaller than that of the data symbols. For example, in the case of six
supportable cyclic time shifts and three (or two) orthogonal time spreading codes in normal (or
extended) CP with three (or two) RS symbols, acknowledgments from 6x3=18 (or 12) different UEs
can be multiplexed within one PUCCH RB. The length-2/4 orthogonal block spreading codes are
based on WalshHadamard codes, and the length-3 spreading codes are based on DFT codes. A
subset of size-s orthogonal spreading codes of a particular length L (s L) is used depending on the
number of RS SC-FDMA symbols. For normal CP with four data SC-FDMA symbols and three
orthogonal time spreading codes (due to there RS symbols), indices 0, 1, 2 of the length-4 orthogonal
spreading codes are used for data time-domain block spreading.
Page 80
Similarly, for extended CP with four data SC-FDMA symbols but only two RS symbols, orthogonal
spreading code indices 0 and 2 of length-4 are used for the data block spreading codes.
The ACK/NACK and SRS may be configured in the same subframe configured by SI. One option is
for the ACK/NACK to take precedence over the SRS, only HARQ is transmitted in the relevant
subframe. The alternative is to configure UEs to use shortened PUCCH in such subframes, whereby
the last SC-FDMA symbol in the second slot of the subframe is not transmitted. This is known as a
shortened PUCCH format, length of the timedomain orthogonal block spreading code is reduced by
one (compared to the first slot). Hence, it uses the length-3 DFT spreading codes in place of length-4
WalshHadamard codes. A UE may not simultaneously transmit on SRS and PUCCH or PUSCH to
avoid violating the single-carrier nature of the signal. Therefore, a PUCCH or PUSCH symbol may be
punctured if SRS is transmitted.
The number of HARQ resource-index
NPUCCH,RB= c.P
where
c=3(normal CP) or 2(extended CP)
P=12/(dPUCCHshift(1,2 or 3)).
As in CQI, cyclic time shift hopping is used for HARQ. In SPS, PDSCH without a DL grant on PDCCH,
PUCCH ACK/NACK resource index n(1)PUCCH should to be used by a UE for initial HARQ. For
dynamically DL scheduling, (including HARQ retransmissions for SPS) on PDSCH (indicated by DL
assignment on PDCCH), PUCCH HARQ resource index n(1)PUCCH is determined based on the index
of the first Control Channel Element (CCE) of DL control assignment. PUCCH region m used for
HARQ with format 1/1a/1b for the case with no mixed PUCCH region is given by
(2)
m = (nPUCCH/NPUCCH,RB) + N RB (RB for PUCCH format 2a/2b).
The PUCCH resource index n(1)(ns ), corresponding to a combination of a cyclic time shift and
orthogonal code (nPUCCH RS and noc), within the PUCCH region m in even slots is given by
n(1)(ns) = n(1)PUCCH mod N(1)PUCCH,RB for ns mod 2 = 0
The PUCCH resources are first indexed in cyclic time shift domain, followed by orthogonal time
spreading code domain. The cyclic time shifts used on adjacent orthogonal codes can also be
staggered, providing the opportunity to separate the channel estimates prior to de-spreading. As high
Doppler breaks down the orthogonality between the spread blocks, offsetting the cyclic time shift
values within each SC-FDMA symbol can restore orthogonality at moderate delay spreads. To
randomize intra-cell interference, PUCCH resource index remapping is used in the second slot. Index
remapping includes both cyclic shift remapping and orthogonal block spreading code remapping. The
PUCCH resource index remapping function in an odd slot is based on the PUCCH resource index in
the even slot of the subframe.
2.2.9 CQI and ACK/NACK in Same RB (Mixed PUCCH RB)

Multiplexing of CQI and HARQ in different PUCCH RBs can in general simplify the system. However,
in small BW as 1.4 MHz control overhead can be high with separate CQI and HARQ RB allocations
(two out of total six). The ZC cyclic time shift structure facilitates the orthogonal multiplexing of CQI
and ACK/NACK signals with different numbers of RS symbols. Assign different sets of adjacent cyclic
time shifts to CQI and ACK/NACK. N(1)cs {0, 1, . . . , 7} cyclic time shifts are used for PUCCH
ACK/NACK formats 1/1a/1b in the mixed PUCCH RB case, whereNcs is integer multiples of dPUCCHshift
. A guard cyclic time shift is used between ACK/NACK and CQI cyclic shift resources to improve
orthogonality and channel separation between UEs transmitting CQI and those transmitting
ACK/NACK. To avoid mixing of the cyclic time shifts for ACK/NACK and CQI, the cyclic time shift (i.e.
the PUCCH resource index remapping function) for the odd slot of the subframe is not used; the same
cyclic time shift as in the first slot of the subframe is used.
2.2.10 Scheduling Request (SR) on PUCCH (Format 1)

The SR PUCCH format 1 structure is same as HARQ PUCCH format 1a/1b, where a cyclic time shift
of base RS sequence is modulated with time-domain orthogonal block spreading. SR uses simple
OnOff keying, with UE transmitting a SR with modulation symbol d(0)=1 to request a PUSCH
resource (positive SR transmission), and transmitting nothing when it does not request to be
scheduled (negative SR).
Page 81
Since HARQ structure is reused for the SR, different PUCCH resource indices (i.e. different cyclic
time shift/orthogonal code combinations) in the same PUCCH region can be assigned for SR (Format
1) or HARQ ACK/NACK (Format 1a/1b) from different UEs. This results in orthogonal multiplexing of
SR and HARQ ACK/NACK in the same PUCCH region. The PUCCH resource index to be used by a
UE for SR transmission, m(1)PUCCH,SRI, is configured by UE-specific signalling. In case a UE needs to
transmit a positive SR in the same subframe as a scheduled CQI transmission, the CQI is dropped
and only the SR is transmitted, in order to maintain the low CM of the transmit signal. Similarly, in the
case of simultaneous SR and SRS configuration, UE does not transmit SRS and transmits only SR. If
an SR and ACK/NACK happen to coincide in the same subframe, UE transmits the ACK/NACK on the
assigned SR PUCCH resource for a positive SR and transmits ACK/NACK on its assigned
ACK/NACK PUCCH resource in case of a negative SR. The modulation mapping is such that a NACK
(or NACK-NACK for 2 codewords) is mapped to +1 resulting in a default NACK in case of DTX similar
to the case of multiplexing CQI and HARQ ACK/NACK for the normal CP.
2.2.11 Control Signalling and UL-SCH multiplexing on

PUSCH
When control signalling is to be transmitted in a subframe in PUSCH, it is multiplexed with UL-SCH
data prior to DFT-spreading; PUCCH is never transmitted in the same subframe as PUSCH. The
CQI/PMI, HARQ ACK/NACK, and RI is multiplexed with the PUSCH data symbols onto uplink
resource elements.
Fig 3.2.11 UCI and PUSCH multiplexing

The RE used for each of CQI/PMI, ACK/NACK and RI is based on the MCS assigned for PUSCH and
an offset parameter, CQIoffset, HARQ-ACKoffset , or RIoffset, configured by higher-layer, different code
rates can be used for control. PUSCH data and control are mapped to different REs. Controls are
mapped to be present in both slots. eNodeB has prior knowledge of UL control signaling, it can easily
de-multiplex control and data. CQI/PMI is placed at the beginning of UL-SCH data RE and mapped
sequentially to all symbols on one subcarrier before continuing on next subcarrier. UL-SCH data is
rate-matched around CQI/PMI data with same MCS. For small CQI and/or PMI report sizes up to 11
bits, a (32,k) block code, similar to the one used for PUCCH, is used, with optional circular repetition
of encoded data; no CRC is applied. For large CQI reporting modes (> 11 bits), 8-bit CRC is attached
and channel coding and rate matching is performed using the tail-biting convolutional code.
ACK/NACK resources are mapped, by puncturing PUSCH data, positions next to the RS, to benefit
from the best possible channel estimation. The maximum RE for ACK/NACK is 4 symbols.
The coded RI are placed next to ACK/NACK symbol positions irrespective of whether ACK/NACK is
actually present in a given subframe. The modulation of ACK/NACK or RI is such that the distance of
symbols carrying ACK/NACK and RI is maximized. The outermost constellation points of the higherorder 16/64-QAM PUSCH modulations are used, resulting in increased transmit power for
ACK/NACK/RI relative to the average PUSCH data power. The coding of RI and CQI/PMI are
separate, with UL-SCH data rate-matched around RI similarly to CQI/PMI. For 1-bit ACK/NACK or RI,
repetition coding is used. For 2-bit ACK/NACK/RI, a (3, 2) simplex code is used with optional circular
repetition of encoded data. The resulting code achieves the theoretical maximum values of the
minimum Hamming distance of the output codewords in an efficient way.
Page 82
Control signalling can also be transmitted on PUSCH without UL-SCH data. CQI/PMI, RI, and/or
ACK/NACK are multiplexed prior to DFT-spreading. ACK/NACK is mapped next to RS, by puncturing
the CQI data and RI symbols, irrespective of whether ACK/NACK is actually present in a given
subframe. The reference CQI/PMI MCS is computed from the CQI payload size and resource
allocation. The channel coding and rate matching of the control signalling without UL-SCH data is the
same as that of multiplexing control with UL-SCH data as described above.
2.2.12 Multiple-Antenna Techniques

Uplink closed-loop antenna selection (for up to four transmit antennas) is supported. If UE signals
MIMO support, eNodeB configures accordingly and schedules the UE. If the eNodeB enables a UEs
closed-loop antenna selection capability, the SRS transmissions then alternate between the transmit
antennas in successive configured SRS transmission subframes, irrespective of whether frequency
hopping is enabled or disabled.
2.2.13 PUSCH UE Antenna Selection Indication

When closed-loop antenna selection is enabled, eNodeB indicates which antenna should be used for
the PUSCH by uplink scheduling grant ( DCI Format 0), the 16 CRC parity bits are scrambled
(modulo-2 addition) by an antenna selection mask. The antenna selection mask is applied in addition
to the UE-ID masking which indicates for which UE the scheduling grant is intended. This implicit
encoding avoids the use of an explicit antenna selection bit which would result in an increased
overhead for UEs not supporting transmit antenna selection. CRC is masked by both the antenna
selection indicator and the 16-bit UEID. The UEID can be detected directly from the 15 least
significant bits of the decoded mask without needing to use the transmitted antenna selection mask
(bit 16). The UE behaviour for adaptive/non-adaptive HARQ retransmissions when configured for
antenna selection is as follows:
1. Adaptive HARQ.
The antenna indicator (via CRC masking) is always sent in the UL grant to indicate which
antenna to use. For example, for a high Doppler UE with adaptive HARQ, the eNodeB might
instruct the UE to alternate between the transmit antennas, or alternatively select the primary
antenna. In typical UE implementations, a transmit antenna gain imbalance of 3 to 6 dB
between the secondary and primary antenna is not uncommon.
2. Non-adaptive HARQ.
The UE behaviour is unspecified as to which antenna to use. Thus, for low Doppler
conditions, the UE could use the same antenna as signalled in UL grant, while at high
Doppler the UE could hop between antennas or just select the primary antenna. The antenna
indicated on the UL grant may not be the best and it is better to let the UE select the antenna
to use. If the eNodeB wishes to instruct the UE to use a specific antenna for the
retransmissions, it can use adaptive HARQ.
2.2.14 Multi-User Virtual MIMO or SDMA

Uplink MU-MIMO, multiple UEs transmit on the same set of RBs, each using a single transmit
antenna. For individual UE, it cant make out the difference, predominantly a matter for the eNodeB to
handle scheduling and uplink reception.
To support uplink MU-MIMO, orthogonal DM RS with different cyclic time shifts are used to enable the
eNB to derive independent channel estimates for UL from each UE.
A cell can assign up to eight different cyclic time shifts using the 3-bit PUSCH cyclic time shift offset
on UL grant. So, up to eight UEs can be supported in a cell on same RB. SDMA between cells (i.e.
uplink inter-cell cooperation) is supported by assigning the same base sequence groups and/or RS
hopping patterns to the different cells.
The main uplink physical channels are PUSCH and PUCCH. PUSCH supports both frequencyselective scheduling and frequency-diverse(hopping) transmissions.
Control signalling (consisting of ACK/NACK, CQI/PMI and RI) is carried by PUCCH when no PUSCH
allocated. PUCCH is deliberately mapped to edge RB, to reduce out-of-band emissions caused by
Page 83
data on the inner RBs, as well as maximizing flexibility for PUSCH scheduling in the central part. In all
cases of multiplexing different kinds of control signalling, single-carrier property is preserved. Control
signalling from multiple UEs is multiplexed via orthogonal coding by using cyclic time shift
orthogonality and/or time-domain block spreading. MIMO is used in UL, in particular through closedloop switched antenna diversity and SDMA. These techniques are also cost-effective for a UE
implementation, as they neither assume simultaneous transmissions from multiple UE antennas.
Page 84
2.3. Uplink Reference Signal

As in DL, SCFDMA uplink incorporates Reference Signals (RSs) for data demodulation and channel
sounding. The roles of UL RSs include channel estimation to aid coherent demodulation, channel
quality estimation for scheduling, power control, timing estimation and direction-of-arrival estimation to
support downlink beam-forming. Two types of RS are supported on the uplink:
1. DeModulation RS (DM RS) - Associated with PUSCH and/or PUCCH. These RSs are
primarily used for channel estimation for coherent demodulation.
2. Sounding RS (SRS) - not associated with data and/or control, and primarily used for channel
quality determination to enable frequency-selective scheduling on UL.
RSs are time-multiplexed with data symbols. The DM RSs of a given UE occupy the same RBs as its
PUSCH/PUCCH data transmission, hence each UE RSs are orthogonal. to each other. The SRSs, if
configured, are transmitted on the last symbol in a subframe; SRS can occupy a bandwidth different
from that used for data transmission. UEs transmitting SRS in the same subframe can be multiplexed
via either FDMA or CDMA.
RS characteristics:
1. Constant amplitude in all allocated subcarriers for unbiased channel estimates.
2. Low Cubic Metric (CM) (at worst no higher than data transmissions).
3. Good autocorrelation for accurate channel estimation.
4. Good cross-correlation properties between different RSs in other (or same) cells.
2.3.1. UL RS Signal Sequence Generation

Uplink RS are based on ZadoffChu (ZC) sequences as they satisfy desirable properties for RS
exhibiting- 0 dB CM, ideal cyclic autocorrelation, and optimal cross-correlation.
cross-correlation - an interfering signal gets spread evenly in the time domain after time-domain
correlation of the received signal with the desired sequence - more reliable detection of channel taps.
CM CM of a ZC sequence is degraded from theoretical 0 dB value at Nyquist sampling rate from
unused guard subcarriers at each end of the sequence (number of occupied RS subcarriers is < IFFT
size of OFDM modulator), and the ZC sequence gets oversampled in the time domain.
RS sequence length, Np, = number of subcarriers, ( multiple of subcarriers per RB, NSC(/RB) = 12)
Np = MSC(/RS) = m NSC(/RB) 1 m NRB(/UL)(uplink system bandwidth in terms of RBs).
The length-Np RS sequence is directly applied (without DFT spreading) to Np RS subcarriers at the
input of IFFT.
ZC sequence of odd-length NZC =aq(n) = exp [ j2q ( n(n + 1)/2 + ln )/ NZC) ]
where q = 1, . . . ,NZC 1 is ZC sequence index (root index), n = 0, 1, . . . , NZC 1, and l = 0 in LTE.
In LTE, NZC = (largest prime number smaller <= Np). ZC sequence of length NZC is then cyclically
extended to the target length Np as: rq(n) = aq(n mod NZC), n = 0, 1, . . . , Np 1
Cyclic extension preserves the constant amplitude and zero autocorrelation cyclic shift orthogonality.
Cyclic extension of ZC sequences is used rather than truncation, as in general it provides better CM
characteristics. For sequence lengths of three or more RBs, this provides at least 30 sequences with
CM smaller than or close to that of QPSK.
30 special RS sequences are defined in LTE for resource allocations of one or two RBs, as there are
very few ZC sequences within 2 RB. These special sequences are QPSK rather than ZC-based
j(n)/4
sequences. The QPSK RS sequences are given by r(n) = e
, n= 0, 1, . . . ,MSC(/RS) 1, where
MSC(/RS) =#subcarriers.
.
2.3.2. Base RS Sequences and Sequence Grouping

Cell supports UL transmissions of different bandwidths. Cell is assigned at least one base RS
sequence for each possible RB allocation size. Multiple RS sequences for each allocation size are
then derived from each base sequence by means of different cyclic time shifts.
Page 85
Smallest number of available base sequences is for three RBs, where, only 30 extended ZC
sequences exist. As a result, complete set of base sequences is divided into 30 non-overlapping
sequence-groups. Cell is then assigned one of the sequence-groups for UL from UEs served by the
cell.
For each RB size up to five RBs, each of the 30 sequence groups contains only one base sequence,
since for five RBs (i.e. sequences of length 60) only 58 extended ZC-sequences are available. For
sequence lengths greater than five RBs, more extended ZC-sequences are available, and therefore
each of the 30 sequence-groups contains two base sequences per resource allocation size; this is
exploited in LTE to support sequence hopping (within the sequence-group) between the two slots of a
subframe.
The base sequences for resource allocations larger than three RBs are selected such that they are
the sequences with high cross-correlation to the single 3 RB base sequence in the sequence-group.
Since cross-correlation between the 3 RB base sequences of different sequence-groups is low due to
the inherent properties of the ZC sequences, such a method for assigning the longer base sequences
to sequence-groups helps to ensure that the cross correlation between sequence-groups is kept low,
thus reducing inter-cell interference.
The v base RS sequences of length 3 RBs or larger (i.e.MRS sc 36) assigned to a sequence-group
u are given by,
ru,v(n) = aq (n mod NZC), n = 0, 1, . . . ,MRS sc 1
where u {0, 1, . . . , 29} is the sequence-group number,
v is the index of the base sequence of length MRSsc within the sequence-group u, given by
v = 0, 1 forMRSsc 72
= 0 otherwise
NZC is the largest prime number smaller than MRSsc , and
q is the root ZC sequence index.
2.3.3. Orthogonal RS via Cyclic Time-Shifts of a Base

Sequence
UEs which are assigned to different sets of subcarriers or RBs, transmit RS signals on these
subcarriers and hence achieve separability of the RSs via FDM. However, in certain cases, UEs can
be assigned to transmit on the same set of subcarriers, for example in the case of uplink multi-user
MIMO (Spatial Division Multiple Access (SDMA) or Virtual MIMO). In these cases RSs can interfere
with each other, and some means of separating the RSs from the different transmitters is required.
Using different base sequences for different UEs transmitting in the same RBs is not ideal due to the
non-zero cross-correlation between the base sequences which can degrade the channel estimation at
the eNodeB. It is preferable that the RS signals from the different UEs are fully orthogonal. In theory,
this could be achieved by FDM of the RSs within the same set of subcarriers, although this would
reduce the RS sequence length and the number of different RS sequences available; this would be
particularly undesirable for low-bandwidth transmissions.
Therefore in LTE, orthogonality between RSs occupying the same set of subcarriers is instead
provided by exploiting the fact that the correlation of a ZC sequence with any Cyclic Shift (CS) of the
same sequence is zero. As the channel impulse response is of finite duration, different transmitters
can use different cyclic time shifts of the same base RS sequence, with the RSs remaining orthogonal
provided that the cyclic shifts are longer than the channel impulse response.
If RS SC-FDMA symbol duration is Tp and channel impulse response duration is less than Tcs, then
up to Tp/Tcs different transmitters can transmit in the same symbol, with different cyclic shift values,
with separable channel estimates at the receiver. For example, if Tp/Tcs = 4 and there are four
transmitters, then each transmitter t {1, . . . , 4} can use a cyclic time shift of (t 1)Tp/4 of the same
base sequence. At the eNodeB receiver, by correlating the received composite signal from the
different transmitters occupying the same set of subcarriers with the base sequence, the channel
estimates from the different transmitters are separable in the time domain.
Page 86
Fig 3.3.3- Cyclic Shift Illustration

Since a cyclic time shift is equivalent to applying a phase ramp in the frequency domain, the
frequency-domain representation of a base sequence with cyclic shift, , is given by
r() u,v (n) = ejn .ru,v(n)

where .ru,v(n) is the base (or unshifted) sequence of sequence-group u, with base sequence index v
within the sequence-group, = 2nt/P with nt the cyclic time shift index for transmitter t, and P is the
number of equally spaced cyclic time shifts supported. In the example, P =4 and n = 0..3 for the four
transmitters respectively.
In LTE, 12 equally spaced cyclic time shifts are defined for DM RS on PUSCH and PUCCH. This
allows for delay spreads up to 5.55 s (total 1 symb duration, 12*5.55=.66.6 s).
Cyclic time shifts spaced the furthest apart experience the least cross-talk between the channel
estimates. Thus, when the number of UEs using different cyclic time shifts is less than the number of
cyclic time shifts supported (P), it is beneficial to assign cyclic time shifts with the largest possible
(circular) separation, and this is supported in LTE. Alternatively, cyclic time shift hopping can be
employed for UE. In LTE, hopping of the cyclic time shifts between the two slots in a subframe is
always enabled.
2.3.4. Sequence-Group Hopping and Planning

LTE supports both RS sequence-group hopping and RS sequence-group planning modes
configurable by RRC. Sequence-group assigned to a cell is a function of its physical-layer cell identity
(PCI) and can be different for PUSCH and PUCCH.
Sequence-Group Hopping
It is enabled in a cell by 1-bit broadcast parameter groupHoppingEnabled. This mode is a
combination of hopping and shifting of the sequence-group according to one of 504 sequence-group
hopping/shifting patterns corresponding to the 504 unique cell-IDs. Since there are 30 base
sequence-groups, 17(= 504/30) unique sequence-group hopping patterns of length 20 are defined
(corresponding to 20 slots in a frame), each of which can be offset by one of 30 sequence-group shift
offsets. The sequence-group number u depends on the sequence-group hopping pattern fgh (0..16) and
the sequence-group shift offset fss. The sequence-group hopping pattern changes u from slot to slot in
a pseudo-random manner, while the shift offset fss is fixed in all slots. Both fgh and fss depend on the
cell-ID.
The pattern fgh is obtained from a length-31 Gold sequence generator, of which the second
constituent M-sequence is initialized at the beginning of each radio frame by fgh of the cell. Up to 30
cell-IDs can have the same fgh (planned coordinated cell cluster), with different fss. The same fgh is
used for PUSCH DM RS, SRS and PUCCH DM RS. The fss can be different for PUSCH and PUCCH.
For PUSCH, plan to assign cell-IDs such that the same fgh and fss, and hence same base sequences,
are used in adjacent cells. This can enable RSs from UEs in adjacent cells to be orthogonal to each
other by using different cyclic time shifts of the same base sequence. Therefore fss for PUSCH is
explicitly configured by 5bit, groupAssignmentPUSCH(dss) such that fss = (cell-ID mod 30 + dss)
mod 30, where dss {0, . . . , 29}.
Page 87
For PUCCH, the same RBs at the edge of the system bandwidth are normally used by all cells. To
randomize interference on PUCCH between neighbouring cells with same fgh, fss is simply given by
cell-ID mod 30. Similarly, for SRS, the same fss as PUCCH is used.
There are two base sequences per sequence-group for each RS sequence length greater than 60 (5
RBs), with possibility of sequence-hopping between two base sequences at the slot boundary in the
middle of each subframe. If sequence-group hopping is used, base sequence automatically changes
between each slot, and therefore additional sequence hopping within the sequence group is not
needed; therefore only the first base sequence in the sequence group is used if sequence-group
hopping is enabled.
Sequence-Group Planning
If sequence-group hopping is disabled, the same sequence-group number u, is used in all slots and u
= fss. Planned sequence-group assignment is possible for up to 30 cells in LTE. In this case,
neighbouring cells to be assigned sequence groups u with low cross-correlation.
The same sequence-group number (base sequences) are used in the three cells of each eNodeB,
with different cyclic time shifts assigned to each cell. With sequence-group planning, sequence
hopping within group between two slots of a subframe is enabled by 1-bit parameter,
sequenceHoppingEnabled. The base sequence index for MRS sc 72 used in slot ns is then
obtained from length-31 Gold sequence generator. In order to enable the use of the same base RS
sequence (and hopping pattern) in adjacent cells for PUSCH, the pseudo-random sequence
generator is initialized at the beginning of each radio frame by the sequence-group hopping pattern
index (based on part of the cell-ID), offset by the PUSCH sequence-group shift index of the cell.
2.3.5. Cyclic Shift Hopping

Cyclic time shift fss hopping is always enabled for UL. For PUSCH with P = 12 evenly-spaced cyclic
time shifts, hopping is between the two slots in a subframe, with the cyclic shift () for a UE being
derived in each slot from a combination of (1) a 3-bit cell broadcast cyclic time shift offset
parameter, (2) a 3-bit cyclic time shift offset indicated in each uplink grant and (3) a pseudorandom cyclic shift offset obtained from output of the length-31 Gold sequence generator.
It should be possible to use different cyclic time shifts in adjacent cells with the same sequence-group
(cells of same eNodeB), in order to support orthogonal RS transmissions from UEs in different cells.
This is similar to initialization of pseudo-random sequence generator for sequence hopping within a
sequence-group, where same hopping patterns are needed in neighbouring cells. Thus, the
initialization of PUSCH DMRS cyclic shift pseudo-random sequence generator is the same as
sequence hopping pattern generator, initialized every radio frame.
For PUCCH, cyclic time shift hopping (among P = 12 evenly spaced cyclic time shifts) is performed
per SC-FDMA symbol, with cyclic shift for a given SC-FDMA symbol in a given slot being derived
from a combination of the assigned PUCCH resource index and the output of the length-31 Gold
sequence generator. In order to randomize interference on the PUCCH arising from the fact that the
same band-edge RBs are used for PUCCH transmissions in all cells, the pseudo-random sequence
generator is initialized at the beginning of each radio frame by the cell-ID.
In addition, to achieve intra-cell interference randomization for the PUCCH DM RS, the cyclic time
shift used in the second slot is hopped such that UEs which are assigned adjacent cyclic time shifts in
the first slot use non-adjacent cyclic time shifts (with large separation) in the second slot. A further
benefit of using a different cyclic time shift in each slot is that the non-ideal cross-correlation between
different base RS sequences is averaged (as the cross-correlation is not constant for all time lags).
2.3.6. Demodulation Reference Signals (DM RS)

The DM RSs present in every UL slot, with UL PUSCH or PUCCH are provided for channel estimation
for coherent demodulation.
RS Symbol Duration
Page 88
For a given RS overhead, RSs could in theory be concentrated in one position in each slot, or divided
up and positioned in multiple locations in each slot:
1. One RS symbol per slot, having the same duration as a data symbol ( Long Block (LB)), with
same subcarrier spacing as data symbols.
2. Two RS symbols per slot, each of half the duration of a data symbol ( Short Block (SB)), with
subcarrier spacing in RS symbols being double that of data symbols (only six subcarriers per
RB in the RS symbols).
The LB RS structure was adopted in LTE for the PUSCH DM RS. The exact position of the single
PUSCH DM RS symbol in each uplink slot depends on whether the normal or extended CP is used.
rd
For normal CP with seven SC-FDMA symbols per slot, the PUSCH DM RS occupies the centre (3 ).
nd
With six SC-FDMA symbols per slot. In extended CP, the 2 SC-FDMA symbol is used. For PUCCH,
the position and number of DM RS depends on the type of PUCCH format.
The DM RS occupies the same RBs as RB allocation for PUSCH or PUCCH. Thus, RS sequence
length, MRSsc = #subcarriers for PUSCH or PUCCH. Since PUSCH RB size is multiples of two, three
and/or five RBs, DM RS sequence lengths are also restricted to the same multiples. For interference
randomization, cyclic time shift hopping is always enabled for DM RS.
2.3.7. Uplink Sounding Reference Signals (SRS)

The SRS, which are not associated with PUSCH and/or PUCCH, are used for channel quality
estimation to enable frequency-selective scheduling on UL or to enhance power control or to support
various start-up functions not recently scheduled. Some examples include initial MCS selection,
power control, timing advance, and frequency semi-selective scheduling in which the frequency
resource is assigned selectively for the first slot of a subframe and hops pseudo-randomly to a
different frequency in the second slot.
SRS Subframe Configuration and Position

SRS subframes within a cell is indicated by 4-bit broadcast parameter srsSubframeConfiguration
indicating 15 possible sets of subframes in which SRS may be transmitted within each radio frame.
This configurability provides flexibility in adjusting the SRS overhead depending on the deployment
scenario. A 16th configuration switches SRS off, for a cell serving primarily high-speed UEs.
SRS transmissions are always in the last SC-FDMA symbol in configured subframes. PUSCH data
transmission is not permitted on the SC-FDMA symbol designated for SRS, resulting in a worst-case
sounding overhead (SRS in every subframe) of 7%.
Duration and Periodicity of SRS Transmissions

eNodeB may either request an individual SRS from a UE or configure UE to transmit SRS periodically
until terminated; a 1-bit UE-specific parameter, duration, indicates periodicity- which may be any of
2, 5, 10, 20, 40, 80, 160 or 320 ms. The SRS periodicity and SRS subframe offset within period are
configured by a 10-bit UE-specific parameter called srsConfigurationIndex.
SRS Symbol Structure

To support frequency-selective scheduling between multiple UEs, it is necessary that SRS should
overlap. Interleaved FDMA is used in SRS symbol, with a RePetition Factor (RPF) of 2, giving the
spacing between occupied subcarriers of an SRS signal with a comb-like spectrum. RPF = 2 implies
that the signal occupies every 2nd subcarrier within the allocated sounding bandwidth. Using a larger
RPF could in theory have provided more flexibility in how the bandwidth could be allocated between
UEs, but it would have reduced the sounding sequence length and the number of available SRS
sequences, and therefore RPF is limited to 2 in LTE.
A UE is assigned, as part of its configurable SRS parameters, the transmissionComb index (0 or 1)
on which to transmit the SRS. RS sequences for SRS are the same as for the DM RS, restricted to
multiples of two, three and/or five times the resource block size. In addition, SRS bandwidth (in RBs)
must be an even number, due to RPF of 2 and the minimum SRS sequence length being 12.
Therefore, the possible SRS bandwidths, NSRS-RB (RBs), and the SRS sequence length,MSRSsc , are
respectively given by,
(1+2)
NSRS RB = 2
3 5
and
MSRS sc =
NSRS RB 12
Page 89
where 2, 3, 5 is a set of positive integers. Simultaneous SRS can be transmitted from multiple UEs
using the same RBs and the same offset of the comb, using different cyclic time shifts of the same
base sequence to achieve orthogonal separation. For SRS, eight (evenly-spaced) cyclic time shifts
per SRS-comb are supported, with the cyclic shift being configured individually for each UE.
SRS Bandwidths
Some of the factors which affect the SRS bandwidth are max power of the UE, number of supportable
sounding UEs, and the sounding bandwidth needed to benefit from uplink channel-dependent
scheduling. Full bandwidth sounding provides the most complete channel information when the UE is
sufficiently close to the eNodeB, but degrades as the path-loss increases when the UE cannot further
increase its transmit power to maintain the transmission across the full bandwidth. Full bandwidth
transmission of SRS also limits the number of simultaneous UEs whose channels can be sounded,
due to the limited number of cyclic time shifts (eight cyclic time shifts per SRS-comb).
To improve the SNR and support a larger number of SRS, up to four SRS bandwidths can be
simultaneously supported in LTE depending on the system bandwidth. To provide flexibility with the
values for the SRS bandwidths, eight sets of four SRS bandwidths are defined for each possible
system bandwidth. RRC signalling indicates which of the eight sets is applicable in the cell by means
of a 3-bit cell-specific parameter srsBandwidthConfiguration. This allows some variability in the
maximum SRS bandwidths, which is important as the SRS region does not include the PUCCH region
near the edges of the system bandwidth, which is itself variable in bandwidth.
SRS BW
SRS-BW SRS-BW SRS-BW SRS-BW
configuration
0
1
2
3
0
48
24
12
4
1
48
16
8
4
2
40
20
4
4
3
36
12
4
4
4
32
16
8
4
5
24
4
4
4
6
20
4
4
4
7
16
4
4
4
The specific SRS bandwidth is configured by a further 2-bit UE-specific parameter, srsBandwidth.
The smallest sounding bandwidth supported is 4 RBs. Frequency hopping can be enabled or disabled
for an individual UE based on frequencyDomainPosition. The tree structure of the SRS bandwidths
limits the possible starting positions for the different SRS bandwidths, reducing the overhead for
signalling the starting position to 5 bits (for each UE by Frequency-domain-position.
UL RS provided in LTE fulfil an important function in facilitating channel estimation and channel
sounding. The ZC-based sequence design can be seen to be a good match to this role, with constant
amplitude in the frequency domain and the ability to provide a large number of sequences with zero or
low correlation. This enables both interference randomization and interference coordination
techniques to be employed in LTE system deployments, as appropriate to the scenario. A high degree
of flexibility is provided for configuring the reference signals, especially for the sounding reference
signals, where the overhead arising from their transmission can be traded off against the
improvements in system efficiency which may be achievable from frequency-selective uplink
scheduling.
Page 90
2.4. Uplink Capacity and Coverage

Lets now look at LTE UL system performance of its different technology enablers and their respective
practical constraints. UL multiple access employs SC-FDMA waveform, also known as DFT-S-OFDM.
This technology enables intra-cell orthogonality between different UEs by Cyclic Prefix (CP), provided
each UEs is adequately time-aligned with its serving cell. Intra-cell orthogonality is the main reason
why a spectral efficiency two to three times higher than that of a WCDMA uplink can be achieved. CP
also facilitates simplified receiver structure with frequency-domain equalization in eNodeB, improving
UL spectral efficiency.
Main drivers for UL performance are summarized Here.
1. Short subframe duration (1 ms), Low HARQ Round-Trip Time, RTT (8 ms) - Reduces
latency and increases number of transmission opportunities
2. Low DMRS overhead- Only 2/14 of subframe resources
3. Low control channel overheadTypically 16/100 of subframe resources
4. MMSE(Minimum Mean Squared Error) equalization- Basic DFT receiver
5. Intra-cell orthogonality (low inter-user interference) - Due to SC-FDMA + CP + time
alignment
6. Power control and interference avoidance - Enable frequency reuse of 1
7. Semi-persistent scheduling - Avoids control channel limitation for services with periodic and
frequent transmission of small packets (e.g. VoIP)
8. FDM resource allocation with fine frequency granularity (180 kHz)- Narrow bandwidth
allocations improve cell edge performance when noise-limited.
For study assumption is taken for 10 MHz bandwidth, 20 dB building penetration loss and 3 km/h UE
speed. Case 1 and Case 3 is assumed for inter-eNodeB distances of 500 m and 1732 m respectively.
2.4.1 Uplink Capacity - Factors Affecting Uplink Capacity

Uplink Frequency Reuse and InterferenceMitigation
A frequency reuse factor of one is used for best spectral efficiency, even though one cell will
experience interference from other closest neighbour cells. To avoid low cell-edge throughput
performance, it is important to employ interference mitigation techniques which allow an efficient
trade-off between cell-edge performance and average spectral efficiency across the whole cell. Such
interference mitigation techniques include:
1. Coordination/avoidance;
2. Inter-cell interference randomization;
3. Frequency-domain spreading;
4. Slow power control.
Uplink cell capacity is constrained by interference levels from other active UEs. The ratio between the
total received power spectral density Io(signal+interference), and thermal noise level N0 at eNodeB
receiver j in the time-frequency region k composed of NULsymb symbols and NULsc subcarriers, is
denoted IoT (j, k) and defined as IoT (j, k) = (Io(j, k) + N0)/N0. Without loss of generality the indices j
and k can be dropped, and IoT represents the level for an arbitrary time-frequency region.
Cell IoT levels must be managed to maintain cell-edge coverage for uplink control channels including
random access channels (vital for scheduling data channels and handovers), as well as maintaining
the minimum cell edge data rates for crucial services such as VoIP.
Average IoT varies versus VoIP loading with two different scheduling methods, A(SPS) and
B(Persistent) with a signalled resource assignment bitmap. As the load increases, it becomes more
likely that the same time and frequency resource regions will be occupied in neighbouring cells, such
that average IoT increases.
Interference Management with Separate Control and Data Regions
PUCCH is mapped to the outer edges of BW, which enables IoT to be separately managed (e.g.
different power control algorithms for PUSCH and PUCCH). CQIs and ACK/NACK are assigned in
Page 91
such a manner as to reduce inter-intra-cell control channel interference under light to moderately
loaded conditions.
2.4.2 Uplink Power Control and Interference Management

Uplink power control optimizes uplink system capacity. Each UE uses path-loss measurements based
on Scells RSs to determine compensation for a fraction of the path-loss. Such fractional power
control is parameterized to effect a tradeoff between overall spectral efficiency and cell edge
performance, with explicit closed-loop power control commands. Power control is combined with RB
allocation strategies which allow interference coordination to enhance cell edge performance and
spectral efficiency.
eNB can schedule UEs with comparable pathloss in adjacent cells to transmit in the same RB. On
average, such a grouping of UEs with similar channel quality in adjacent cells results in the best cell
edge performance, since it avoids strong interference from UEs close to eNodeB in adjacent cells.
Conversely, aligning UEs with different channel quality between cells will benefit the UEs with good
channel quality, and hence improve peak data rates and average cell throughput. Residual uplink
interference after such interference coordination techniques may be mitigated by receiver techniques
at the eNodeB, for example employing multiple receive antennas for beamforming, or using an
Interference Rejection Combining (IRC) receiver.
.
2.4.3 Uplink Control Channel Overhead

There is a spectral efficiency trade-off between PUSCH and PUCCH. PUCCH needed for control
signalling depends on error rate requirements (typically below 1% for control signalling), the size of
the data packets (more overhead for small packets), and time considered acceptable to switch from
idle to active states. Data transmissions can make use of resources left over after control signalling
allocation. Hence minimizing control signalling is key to maximizing data spectral efficiency.
2.4.4 Modulation and Number of HARQ Transmissions

Uplink data capacity is limited by maximum available modulation order (64QAM) uplink, the number of
receive antennas at eNodeB and the scheduling algorithm. Algorithm can trade off capacity against
fairness and cell edge throughput by appropriate configuration of the power control and RB allocation.
Any QoS constraints (usually a delay limit) will affect uplink capacity.
By enabling 64QAM, the peak rate is increased to 14.3 Mbps for (from 9.6 Mbps for 16QAM) UEs
closest to their serving cell. The VoIP normally is recommended with Transmission Time Interval (TTI)
bundling with a 244 bit full-rate AMR codec and a smaller RTP/UDP/RLC header.
2.4.5 Delay Constraints and VoIP

HARQ RTT (time between retransmissions) and the dropped packet delay bound (i.e. maximum time
a packet will be maintained in the transmitter queue before it is dropped) also affect capacity for
delay-sensitive services like VoIP or other semi-real-time services. For a delay bound of 50 ms
(typical for VoIP), UL HARQ RTT of 8 ms means that up to six transmissions per voice packet are
possible. Packet segmentation and TTI bundling (use of multiple HARQ processes per packet) may
be used to improve coverage.
Best-effort data services (HTTP and FTP traffic) are not subject to tight delay constraints like VoIP,
and can therefore trade off delay for higher throughput, taking advantage of changes in instantaneous
channel conditions through appropriate scheduling.
2.4.6 Number of eNodeB Receive Antennas

Increasing #receive antennas from two to four in each cell may can result in a 50% throughput
improvement given full buffer traffic, even with single transmit antenna UEs. This improvement arises
from the two-fold increase in energy obtained with four-antenna diversity compared to two antennas,
as well as a diversity gain in fading channels.
Page 92
2.4.7 Minimum Size of Resource Allocation

In high Signal-to-Interference plus Noise Ratio (SINR) conditions, maximum achievable capacity can
be limited by the minimum amount of RB which can be allocated to a single UE.
For very high SINR scenarios, the supportable VoIP users in a cell reaches a plateau in case of 1 ms
TTI, as the minimum RB allocation is larger than necessary to transmit a single VoIP packet. To
optimize this a RB is limited to 12 subcarrier. Higher subcarrier in a RB will give opposite effect of the
benefits of TTI bundling, whereby a long TTI has the potential to increase the received energy per
packet, and therefore to improve coverage at the cell edge when the UE is power-limited. The choice
of a 1 ms TTI length in LTE is therefore a compromise between high capacity in high-SINR
conditions and good coverage at the cell edge. A choice of a small RB size (only 12 subcarriers)
helps to ensure that the minimum resource allocation size does not unduly limit capacity.
2.4.8 LTE Uplink Capacity Evaluation

The VoIP capacity is typically quoted as number of users per cell. VoIP in uplink, benefit from HARQ,
which compensates for soft handover. For coverage-limited UEs, a 12 subcarrier RB, helps to
improve coverage by increasing power per subcarrier.
2.4.9 LTE Uplink Coverage and Link Budget

A Wireless Network requires not only good cell-edge performance, but also well balanced channels in
terms of coverage they each provide. Below is typical performance requirement data. This depends
on many factors including cell size and UE distance, shadow margin, Cell coverage %tile etc.
Required UL Es/N0 for target error rate, 5 MHz LTE carrier.
Uplink control channel
-----------------------------ACK/NACK PUCCH
CQI PUCCH
PRACH
Required Es/N0 for Target ER Reference

-------------------------------------------------Es/N0 =7.5 dB
P(NACKACK) = 104
P(ACKNACK) = 102
P(DTXACK) = 102
Es/N0 =7.5 dB (5-bit) for 1% BLER
Es/N0 =4.5 dB (10-bit) for 1% BLER
Es/N0 =13.5 dB for 1% PER
BER targets for UL ACK/NACK signalling for LTE.

Event
--------------------------ACK missed detection
DTX to ACK error
NACK to ACK error
CQI block error rate
Target quality
-----------------102
102
104
Likely to be around 102101
The link budget results are for deployment scenario Case 3, and assume a log-normal
shadowing margin of 12.1 dB corresponding to 98-percentile single-cell area coverage
reliability and a propagation model given by
Propagation loss = 128.1 + 37.6 log(distance(m)) (18.7)
Only an average throughput of 5 kbps can be supported based on a single (1 ms) RB PUSCH
transmission. The downlink SINR at 1000m given the corresponding downlink conditions is about 8.3
dB, which is also the SINR required for a 1% BLER on the Physical Broadcast CHannel (PBCH)
SINR. At 1000 m the transmission loss (propagation loss minus the antenna gains plus the log-normal
shadowing margin plus the penetration/body loss) is 146.2 dB, which must be supported by the uplink
and downlink control channels in order to achieve the 98-percentile area coverage reliability.
1. 1-bit ACK/NACK and 4-bit CQI Coverage via PUCCH.
Page 93
The 1-bit ACK/NACK and the 4-bit CQI, each transported on a PUCCH, achieve almost the
same coverage as the PBCH, but fall short by approximately 1.3 dB as determined by the
difference in transmission loss (146.2144.9 dB). A repetition factor of 2 may be used to close
the coverage gap.
2. 12.2 kbps AMR VoIP Coverage via PUSCH.
Even TTI bundling over four subframes (i.e. allocating four HARQ processes so that each
VoIP packet can occupy four consecutive TTIs) combined with five HARQ transmissions per
packet is not enough to close the coverage gap for a VoIP 12.2 kbps AMR service. Some
relaxation in error rate or reduction of the AMR codec rate (e.g. 7.95 kbps) may be enough to
close the coverage gap.
3. PRACH coverage.
A repeated RACH preamble burst (2 800 s) is needed for the PRACH to achieve 98percentile or better area coverage reliability, since PRACH format 2, only supports a cell
radius of about 0.8 km which is close to the range supported by one ZadoffChu root
sequence (0.78 km). With a single sequence a total received preamble energy per sequence
of approximately 18 dB (Es/N0 11.5 dB) is required to meet missed detection and false
alarm probabilities of less than 1%. PRACH format 2 with repetition is slightly better with a
required Es/N0 =13.5 dB for the same error probabilities.
We highlight the main factors affecting uplink capacity and coverage. It is possible to observe how
uplink coverage issues have led to particular design choices in LTE, such as the size of a RB, the
length of TTI, and the design of the control channels. Evaluations can be carried out to examine the
effect of each relevant factor, leading to the LTE performance being able to be characterized by a
variety of metrics. Such metrics include average throughput, cell-edge throughput, number of VoIP
users, and FTP download time. Depending on the requirements of particular deployments, the
eNodeB has the freedom to balance average throughput against cell-edge coverage and fairness.
Page 94
2.5. Random Access

RACH is used for initial network access, but not to carry any user data, which is exclusively sent on
PUSCH. An UE can only be scheduled for UL, if its UL timing is synchronized. RACH plays this key
role of synchronizing UEs to radio access. Once uplink synchronization is achieved for a UE, the
eNodeB can schedule PUSCH for it. Relevant scenarios in which the RACH is used are therefore:
(1) A UE in RRC_CONNECTED, but not uplink-synchronized, needing to send new uplink data
or control information (e.g. an event-triggered measurement report);
(2) A UE in RRC_CONNECTED, but not uplink-synchronized, needing to receive new downlink
data, and therefore to transmit corresponding ACK/NACK in UL;
(3) A UE in RRC_CONNECTED, handing over from its serving cell to a target cell;
(4) A transition from RRC_IDLE state to RRC_CONNECTED, for example for initial access or
tracking area updates;
(5) Recovering from radio link failure.
(6) an uplink-synchronized UE is to send a Scheduling Request (SR) if it does not have any UL
resource allocated in which to send SR.
These roles require the LTE RACH to be designed for low latency, as well as good detection
probability at low Signal-to-Noise (SNR) in order to guarantee similar coverage to that of PUSCH and
PUCCH.
A successful RACH attempt should allow subsequent UE transmissions to be inserted
among the scheduled synchronized transmissions of other UEs. This sets the required timing
estimation accuracy which must be achievable from the RACH, and hence the required RACH
transmission bandwidth: due to the Cyclic Prefix (CP), the LTE RACH only needs to allow for roundtrip delay estimation (instead of the timing of individual channel taps), and this therefore reduces the
required RACH bandwidth compared to WCDMA.
This is beneficial in minimizing the overhead of the RACH, which is another key consideration. Unlike
in WCDMA, the RACH should be able to be fitted into the orthogonal
time-frequency structure of the uplink, so that an eNodeB which wants to avoid interference between
the RACH and scheduled PUSCH/PUCCH transmissions can do so. It is also important that the
RACH is designed so as to minimize interference generated to adjacent scheduled PUSCH/PUCCH
transmissions.
2.6.1 Random Access Procedure

Random Access (RA) procedure comes in two forms, either contention-based (with risk of collision) or
contention-free.
In contention-based procedure, a RA preamble signature is randomly chosen by the UE, it is possible
for more than one UE simultaneously transmit the same signature, needing subsequent contention
resolution process.
For use-cases (2 - new downlink data) and (3- handover), the eNB has the option of preventing
contention by allocating a dedicated signature to a UE, resulting in contention-free access. This is
faster than contention-based access a factor which is particularly important for the case of
handover, which is time-critical.
Unique 64 preamble signatures are generated by cells broadcast ROOT-ZADOOF-CHU sequence by
UE and eNB in each cell. The operation of the two types of RACH procedure depends on a
partitioning of these signatures between those for contention-based access and those reserved for
allocation to specific UEs on a contention-free basis.
2.6.2 Contention-Based Random Access Procedure

The contention-based procedure consists of four-steps:
Step 1: Preamble transmission;
Step 2: Random access response;
Step 3: Layer 2 / Layer 3 (L2/L3) message;
Step 4: Contention resolution message.
Step 1: Preamble Transmission
.UE selects one of the 64 Ncf available PRACH contention-based preambles, where Ncf is
signatures reserved by eNB for contention-free RACH. The set of contention-based signatures is
Page 95
further subdivided into two subgroups, so that the choice of signature can carry one bit of information
to indicate information relating to the amount of transmission resource needed to transmit the
message at Step 3. The SI indicates which signatures are in each of the two subgroups (each
subgroup corresponding to one value of the one bit of information), as well as the meaning of each
subgroup. The UE selects a signature from the subgroup corresponding to the size of transmission
resource needed for the appropriate RACH use case (some use cases require only a few bits to be
transmitted at Step 3, so choosing the small message size avoids allocating unnecessary UL
resources), which may also take into account the observed DL radio channel conditions. The eNB can
control the number of signatures in each subgroup according to the observed loads in each group.
Fig 3.5.2.1 Preamble formats

The initial preamble Tx power setting is based on open-loop estimation with full compensation for the
path-loss, so that received power of the preambles is independent of the path-loss; designed to help
the eNB to detect several simultaneous preamble Tx in the same PRACH RB resource.
UE estimates the path-loss by DL RSRP. The eNodeB may also configure an additional power offset.
Step 2: Random Access Response
The Random Access Response (RAR) is sent by the eNodeB on PDSCH, and addressed with an ID,
Random Access - Radio Network Temporary Identifier (RA-RNTI), identifying the time-frequency slot
in which the preamble was detected. If multiple UEs had collided by selecting the same signature in
the same preamble subframe, they would each receive the RAR.
RAR conveys the identity of detected preamble, a timing alignment instruction to synchronize UE with
uplink, an initial UL grant for transmission of the Step 3 message, and an assignment of a Temporary
Cell Radio Network Temporary Identifier (C-RNTI) (which may or may not be made permanent as a
result of the next step contention resolution). The RAR message can also include a backoff
indicator which the eNodeB can set to instruct the UE to back off for a period of time before retrying a
RA attempt. The UE expects to receive the RAR within a time window, of which the start and end are
configured by the eNodeB and broadcast as SI. The RAR window starts 2 ms after the end of the
preamble subframe. However, a typical delay (measured from the end of the preamble subframe to
the beginning of the first subframe of RAR window) is more likely to be 4 ms.
Page 96
Fig 3.5.2.2 RACH steps

If UE does not receive a RAR within the configured time window, it retransmits the preamble. The
minimum delay for preamble retransmission after the end of the RAR window is 3 ms. (If the UE
receives the PDCCH signalling the downlink resource used for the RAR but cannot satisfactorily
decode the RAR message itself, the minimum delay before preamble re-transmission is increased to
4 ms, to allow for the time taken by the UE in attempting to decode the RAR.)
Fig 3.5.2.3 RAR (Random Access Response format)

The eNodeB may configure preamble power ramping so that the transmission power for each
retransmitted preamble is increased by a fixed step. RA preambles are normally orthogonal to other
uplink transmissions. Therefore, the proportion of RA attempts which succeed at the first preamble
transmission is likely to be higher than in WCDMA, and the need for power ramping is likely to be
reduced.
Step 3: Layer 2/Layer 3 (L2/L3) Message
This message is the first scheduled UL Tx on PUSCH and uses HARQ. It conveys actual RA
procedure
message, such as an
RRCConnectionRequest, TrackingAreaUpdate, or
SchedulingRequest. It includes UE-id scrambled with Temp C-RNTI allocated in the RAR at Step 2. In
case of a preamble collision, colliding UEs will receive the same Temp C-RNTI through the RAR and
will also collide in the same uplink RB when transmitting their L2/L3 message. This may result in such
interference that no colliding UE can be decoded, and the UEs restart the random access procedure
after reaching the maximum number of HARQ retransmissions. However, if one UE is successfully
decoded, the contention remains unresolved for the other UEs. The following DL message (in Step 4)
allows a quick resolution of this contention.
Page 97
Fig 3.5.2.4 Message sending after RAR

Step 4: Contention Resolution Message
The contention resolution message with specified UE-id is addressed to the C-RNTI or Temporary CRNTI, and, in the latter case, echoes the UE identity contained in the L2/L3 message. It supports
HARQ. In case of a collision followed by successful decoding of the L2/L3 message, the HARQ
feedback is transmitted only by the UE which detects its own UE identity (or C-RNTI); other UEs
understand there was a collision, transmit no HARQ feedback, and can quickly exit the current RA
procedure and start another one. The UEs behaviour upon reception of the contention resolution
message therefore has three possibilities:
1. UE correctly decodes the message and detects its own identity: it sends back a positive
ACKnowledgement, ACK.
2. UE correctly decodes the message and discovers that it contains another UEs identity
(contention resolution): it sends nothing back (DTX).
3. UE fails to decode the message or misses DL grant: it sends nothing back (DTX).
Contention-Free Random Access Procedure

The slightly unpredictable latency of RA procedure can be circumvented when low latency is required,
such as handover and resumption of DL traffic for a UE, by allocating a dedicated signature to the UE
on a per-need basis. In this case the procedure is simplified. The procedure terminates with the RAR.
2.6.3 Physical Random Access Channel Design

The RA preamble part of the RA procedure is mapped onto the PRACH. The PRACH is time- and
frequency-multiplexed with PUSCH and PUCCH. PRACH time-frequency resources are semistatically allocated within the PUSCH region, and repeat periodically. The possibility of scheduling
PUSCH transmissions within PRACH slots is left to the eNodeBs discretion.
The PRACH Structure

DFT-S-OFDM PRACH Preamble Symbol
LTE PRACH preamble consists of a complex sequence. It is also an OFDMsymbol, built with a CP,
thus allowing for an efficient frequency-domain receiver at the eNodeB. The end of the sequence is
appended at the start of the preamble, thus allowing a periodic correlation at the PRACH receiver.
The UE aligns the start of the RA preamble with the start of the corresponding UL subframe at the UE
assuming a timing advance of zero, and the preamble length is shorter than PRACH slot in order to
provide room for a Guard Time (GT) to absorb the propagation delay. As for a conventional OFDM
symbol, a single observation interval can be used regardless of the UEs delay, within which periodic
correlation is possible. The PRACH preamble sequence is optimized with respect to its periodic
autocorrelation property.
PRACH Formats
Page 98
Four Random Access (RA) preamble formats are defined for FDD. Each format is defined by the
durations of the sequence and its CP.
format
---0
TCP (s)
----------103.13
TSEQ (s)
------800
684.38
800
203.13
1600
684.38
1600
14.6
133
usage
---------------------------Normal 1 ms RA burst with 800 s
preamble sequence, for smallmedium cells (up to14 km)
2 ms RA burst with 800 s preamble
sequence, for large cells (up to 77 km) without a link budget problem
2 ms RA burst with 1600 s
preamble sequence,for medium cells (up to 29 km)
supporting low data rates
3 ms RA burst with 1600 s preamble sequence,
for very large cells (up to 100 km)
2 symbol period (approx)
Fig 3.5.3.1 Preable format versus Cell Radius Max and Recommended
Sequence Duration
The sequence duration, TSEQ, is driven by the following factors:
1. Trade-off between sequence length and overhead: a single sequence must be as long as
possible to maximize the number of orthogonal preambles, while still fitting within a single
subframe to keep PRACH overhead small;
2. Compatibility with the maximum expected round-trip delay;
3. Compatibility between PRACH and PUSCH subcarrier spacings;
4. Coverage performance.
Maximum round-trip time.
The lower bound for TSEQ must allow for unambiguous round-trip time estimation for a UE located at
the edge of the largest expected cell (i.e. 100 km radius), including the maximum delay spread
3
8
6
expected in such large cells, namely 16.67 s. Hence TSEQ (200x10 ) / (3x10 )+ 16.67x10 =
683.33 s. For format 3, GT=0.72ms which is sufficient for 103km.
Subcarrier spacing compatibility.
Further constraints on TSEQ are given by SC-FDMA signal generation principle, such that the size of
the DFT and IDFT, NDFT, must be an integer number: NDFT = fs *TSEQ = k, k N, where fs is the
system sampling rate (30.72 MHz=1/Ts). Additionally, it is desirable to minimize the orthogonality loss
in the frequency domain between the preamble subcarriers and the subcarriers of the surrounding
uplink data transmissions. This is achieved if the PUSCH data symbol subcarrier spacing df is an
integer multiple of the PRACH subcarrier spacing dfRA:
dfRA = fs / NDFT = 1/TSEQ = 1/ (k*TSYM) = (1/ k)* df, k N
where TSYM = 66.67 s is the symbol duration. In other words, the preamble duration must be an
integer multiple of UL subframe symbol duration:
TSEQ = kTSYM = k/ df , k N. Here k=12 makes Tseq=12*66.67 = 800s.
The FFT/IFFT components can be reused from SC-FDMA processing for the scheduled data. For
example, an n 2m DFT can be implemented with an FFT of 2m samples combined with a DFT of n
samples, since NDFT = kfsTSYM = kNFFT, k N where NFFT is the FFT size for a PUSCH symbol.
Page 99
Coverage performance.
In general a longer sequence gives better coverage, but better coverage requires a longer CP and GT
in order to absorb the corresponding round-trip delay. The required CP and GT lengths for PRACH
format 0, for example, can be estimated from the maximum round-trip delay achievable by a preamble
sequence which can fit into a 1 ms subframe.
Under a noise-limited scenario (low density cell), coverage performance can be estimated from a link
budget calculation. Under the assumption of the Okumura-Hata empirical model of distancedependent path-loss L(r) (where r is the cell radius in km), the PRACH signal power PRA received at
the eNodeB baseband input can be computed as follows:
PRA(r) = Pmax + Ga L(r) LF PL (dB)
Generic Parameters:
Parameter
Value
---------------------Carrier frequency (f )
2000 MHz
Antenna height (hb)
30 m / 60 m
UE antenna height (hm)
1.5 m
UE transmitter EIRPa (Pmax)
24 dBm (250 mW)
eNodeB Receiver Antenna Gain (Ga)
14dBi
Receiver noise figure (Nf)
5.0 dB
Thermal noise density (N0)
174 dBm/Hz
Percentage of area covered by buildings ()
10%
Required Ep/N0 (eNodeB with 2 Rx antenna)
18 dB (six-path Typical Urban
channel model)
Penetration loss (PL)
0dB
Log-normal fade margin (LF)
0 dB
PRACH preamble sequence duration TSEQ is then derived from the required preamble sequence
energy to thermal noise ratio Ep/N0 to meet a target missed detection and false alarm probability, as
follows:
TSEQ = N0Nf * Ep / (PRA(r) N0)
where N0 is the thermal noise power density (in mW/Hz) and Nf is the receiver noise figure (in linear
scale). Assuming that Ep/N0 = 18 dB is required to meet missed detection and false alarm
probabilities of 102 to 103, It is observed that the potential coverage performance of a 1 ms PRACH
preamble is in the region of 14 km. As a consequence, the required CP and GT lengths are
approximately (2 14000)/(3 108) = 93.3 s, so that the upper bound for TSEQ is given by TSEQ
1000 2 93.33 = 813 s.
Therefore, the longest sequence is TSEQ = 800 s, as used for preamble formats 0 and 1.
The resulting PRACH subcarrier spacing is dfRA = 1/TSEQ = 1.25 kHz. The 1600 s preamble
sequence of formats 2 and 3 is implemented by repeating the baseline 800 s preamble sequence.
These formats can provide up to 3 dB link budget improvement, which is useful in large cells and/or to
balance PUSCH/PUCCH and PRACH coverage at low data rates.
CP and GT Duration
Having chosen TSEQ, the CP and GT dimensioning can be specified more precisely. For formats 0
and 2, the CP is dimensioned to maximize the coverage, given a maximum delay spread d: TCP =
(1000 800)/2 + d/2 s, with d 5.2 s (corresponding to the longest normal CP of a PUSCH
symbol). The maximum delay spread is used as a guard period at the end of CP, thus providing
protection against multipath interference even for the cell-edge UEs.
For a cell-edge UE, the delay spread energy at the end of the preamble is replicated at the end of the
CP and is therefore within the observation interval. Consequently, there is no need to include the
maximum delay spread in the GT dimensioning. Hence, instead of locating the sequence in the centre
of the PRACH slot, it is shifted later by half the maximum delay spread, allowing the maximum
Round-Trip Delay (RTD) to be increased by the same amount. The residual delay spread at the end
of the preamble from a cell-edge UE spills over into the next subframe, but this is taken care of by the
CP at the start of the next subframe to avoid any inter-symbol interference.
For formats 1 and 3, the CP is dimensioned to address the maximum cell range in LTE, 100 km, with
a maximum delay spread of d 16.67 s. In practice, format 1 is expected to be used with a 3LTE Physical Layer- 3PCA-L1 Certification
Page 100
subframe PRACH slot; the available GT in 2 subframes can only address a 77 km cell range. It was
chosen to use the same CP length for both format 1 and format 3 for implementation simplicity. Of
course, handling larger cell sizes than 100 km with suboptimal
CP dimensioning is still possible and is left to implementation. The CP lengths are designed to be an
integer multiple of the assumed system sampling period for LTE, TS = 1/30.72 s.
PRACH Resource Configurations
The PRACH slots can be configured to occur in up to 64 different layouts, or resource configurations.
Depending on the RACH load, one or more PRACH resources may need to be allocated per PRACH
slot period. The eNB has to process the PRACH very quickly so that message 2 of the RACH
procedure can be sent within the required window. In case of more than one PRACH resource per
PRACH period it is generally preferable to multiplex the PRACH resources in time rather than in
frequency. This helps to avoid processing peaks at the eNodeB.
Fig 3.5.3.2 PRACH Config Index Table for FDD

Extending this principle, the available slot configurations are designed to facilitate a PRACH receiver
which may be used for multiple cells of an eNodeB, assuming a periodic pattern with period 10 ms or
20 ms.
Assuming an operating collision probability per UE, pcoll = 1%, one PRACH resource (with 64
signatures) per 10 ms per 5 MHz can handle an offered load G=64 ln(1 pcoll) = 0.6432 average
PRACH attempts, which translates into 128 attempts per second in 10 MHz. This is expected to be a
typical PRACH load in LTE.
Assuming a typical PRACH load, Resource configurations 02 and 15 use a 20 ms PRACH period,
which can be desirable for small BW (1.4 MHz) in order to reduce the PRACH overhead at the price
of higher waiting times.
Page 101
Fig 3.5.3.3 PRACHConfigIndex=3

In a three-cell scenario, time collision of PRACH resources can always be avoided except for the 20
MHz case, where collisions can be minimized to two PRACH resources occurring in the same
subframe. It should also be noted that in a scenario with six cells per eNodeB, at most two PRACH
resources will occur in the same subframe for bandwidths below 20 MHz.
Fig 3.5.3.4 Assign different PRACHConfigIndex to neighbour cells

The variety of configurations provided therefore enables efficient dimensioning of a multicell PRACH
receiver.
2.6.4 Preamble Sequence Theory and Design

As noted above, 64 PRACH signatures are available in LTE, compared to only 16 in WCDMA. This
can not only reduce the collision probability, but also allow for 1 bit of information to be carried by the
preamble and some signatures to be reserved for contention free access. In LTE prime-length
ZadoffChu (ZC) sequences are chosen. These sequences enable improved PRACH preamble
detection performance.
1. The power delay profile is built from periodic correlation;
2. The intra-cell interference between different preambles received in the same PRACH
resource is reduced;
3. Intra-cell interference is optimized with respect to cell size: the smaller the cell size, larger the
number of orthogonal signatures and the better the detection performance;
4. The eNodeB complexity is reduced;
Page 102
5. The support for high-speed UEs is improved.

The 800 s LTE PRACH sequence is built from cyclicly-shifting a ZC sequence of prime-length NZC,
and the sequence length NZC = 839 for FDD.
Preamble Bandwidth
To ease multiplexing of PRACH and PUSCH RB, a PRACH slot must be allocated a bandwidth
BWPRACH equal to an integer multiple of Resource Blocks (RBs), i.e. an integer multiple of 180 kHz.
For simplicity, BWPRACH in LTE is constant for all BW; to optimize both detection performance and
timing estimation accuracy. Timing estimation drives the lower bound of the PRACH bandwidth.
Indeed, a minimum bandwidth of approx 1 MHz is necessary to provide a one-shot accuracy of about
0.5 s, which is an acceptable timing accuracy for PUCCH/PUSCH.
Fig 3.5.4 RACH resource illustration

One expects that higher the bandwidth, better the detection performance, due to diversity gain.
However, it is important to make the comparison using a constant signal energy to noise ratio, Ep/N0,
resulting from accumulation (or despreading) over the same preamble duration, and the same false
alarm probability, pfa, for all bandwidths. The latter requires the detection threshold to be adjusted
with respect to the search window size, which increases with the bandwidth. Indeed, the larger the
search window( higher pfa), the larger the bandwidth the higher the threshold relative to the noise
floor, given a false alarm target pfa_target and cell size L; equivalently, the larger the cell size, the
higher the threshold relative to the noise floor, given a target pfa_target and bandwidth. As a result, a
smaller bandwidth will perform better than a large bandwidth in a single-path static AWGN channel,
given that no diversity improvement is to be expected from such a channel.
We have observed that the best detection performance is achieved by preambles of 6 RBs and 12
RBs for low and high SNRs respectively. The 25-RB preamble has the overall best performance
considering the whole SNR range. Thus the diversity gain of large bandwidths only compensates the
increased detection threshold in the high SNR region corresponding to misdetection performances in
the range of 103 and below. At a typical 102 detection probability target, the 6-RB allocation only has
0.5 dB degradation with respect to the best case.
Therefore, a PRACH allocation of 6 RBs provides a good trade-off between PRACH overhead,
detection performance and timing estimation accuracy. Note that for the smallest system bandwidth
(1.4 MHz, 6 RBs) the PRACH overlaps with the PUCCH; it is left to the eNodeB implementation
whether to implement scheduling restrictions during PRACH slots to avoid collisions, or to let PRACH
collide with PUCCH and handle the resulting interference. Finally, the exact preamble transmission
bandwidth is adjusted to isolate PRACH slots from surrounding PUSCH/PUCCH allocations through
guard bands, as elaborated in the following section.
Sequence Length
The sequence length design should address the following requirements:
1. Maximize the number of ZC sequences with optimal cross-correlation properties;
2. Minimize the interference to/from the surrounding scheduled data on the PUSCH.
Page 103
The former requirement is guaranteed by choosing a prime-length sequence. For the latter, since data
and preamble OFDM symbols are neither aligned nor have the same durations, strict orthogonality
cannot be achieved. At least, fixing the preamble duration to an integer multiple of the PUSCH symbol
provides some compatibility between preamble and PUSCH subcarriers. However, with the 800 s
duration, the corresponding sequence length would be 864, which does not meet the prime number
requirement. Therefore, shortening the preamble to a prime length slightly increases the interference
between PUSCH and PRACH by slightly decreasing the preamble sampling rate.
The interference from PUSCH to PRACH is further amplified by the fact that the operating Es/N0 of
PUSCH (where Es is the PUSCH symbol energy) is much greater than that of the PRACH (typically as
much as 24 dB greater if we assume 13 dB Es/N0 for 16QAM PUSCH, while the equivalent ratio for
the PRACH would be 11 dB assuming Ep/N0 = 18 dB and adjusting by 10 log10 (864) to account for
the sequence length).
The PRACH uses guard bands to avoid the data interference at preamble edges. A cautious design of
preamble sequence length not only retains a high inherent processing gain, but also allows avoidance
of strong data interference. In addition, the loss of spectral efficiency (by reservation of guard
subcarriers) can also be well controlled at a fine granularity (dfRA = 1.25 kHz).
In the absence of interference, there is no significant performance difference between sequences of
similar prime length. In the presence of interference, it can be seen that reducing the sequence length
below 839 gives no further improvement in detection rate. No effect is observed on the false alarm
rate.
Therefore the sequence length of 839 is selected for LTE PRACH, corresponding to 69.91 PUSCH
subcarriers in each symbol, and offers 72 69.91 = 2.09 PUSCH subcarriers protection, which is very
close to one PUSCH subcarrier protection on each side of the preamble. Note that the preamble is
positioned centrally in the block of 864 available PRACH subcarriers, with 12.5 null subcarriers on
each side.
The location in the frequency domain is controlled by the parameter nRAPRB,expressed as a RB
number configured by higher layers and fulfilling 0 nRAPRB NULRB 6.
Cyclic Shift Dimensioning (NCS) for Normal Cells
Sequences obtained from cyclic shifts of different ZC sequences are not orthogonal. Therefore,
orthogonal sequences obtained by cyclically shifting a single root sequence should be favoured over
non-orthogonal sequences; additional ZC root sequences should be used only when the required
number of sequences (64) cannot be generated by cyclic shifts of a single root sequence. The cyclic
shift dimensioning is therefore very important in the RACH design.
The cyclic shift offset NCS is dimensioned so that the Zero Correlation Zone (ZCZ) of the sequences
guarantees the orthogonality of the PRACH sequences regardless of the delay spread and time
uncertainty of the UEs. The minimum value of NCS should therefore be the smallest integer number of
sequence sample periods that is greater than the maximum delay spread and time uncertainty of an
uplink non-synchronized UE, plus some additional guard samples provisioned for the spill-over of the
pulse shaping filter envelope present in the PRACH receiver.
Larger the cell, the larger the cyclic shift required to generate orthogonal sequences, and
consequently, the larger the number of ZC root sequences necessary to provide the 64 required
preambles. The relationship between cell size and the required number of ZC root sequences allows
for some system optimization. In general, the eNodeB should configure NCS independently in each
cell, because the expected inter-cell interference and load (user density) increases as cell size
decreases; therefore smaller cells need more protection from co-preamble interference than larger
cells.
Some examples may show different scenarios with different numbers of parameters such as:
(Number of cyclic shifts per ZC sequence * Number of ZC sequences) >= 64, Cyclic shift size Ncs, Cell Radius
etc. For each scenario, the total number of sequences is 64, but resulting from different combinations of
the number of root sequences and cyclic shifts.

There can be some of the following cases:
1. Case 1: Only one UE transmits a preamble;
2. Case 2: Two UEs transmit a preamble, and the two preamble sequences are generated from
the same root ZC sequence;
Page 104
3. Case 3: Two UEs transmit a preamble, and the two preamble sequences are generated from
different root ZC sequences.
It is observed that, when two preambles are transmitted which are cyclic shifts of the same root
sequence (Case 2), the performance does not degrade compared to the case of only one preamble
being transmitted, confirming the ZCZ property of the shifted sequences. By contrast, when the two
preambles are generated from different root sequences (Case 3), a degradation of 0.250.4 dB is
observed at 1% 0.1% missed detection rates.
NCS set design.
Given the sequence length of 839, allowing full flexibility in signalling NCS would lead to broadcasting
a 10-bit parameter, which is over-dimensioning. As a result, in LTE the allowed values of NCS are
quantized to a predefined set of just 16 configurations. The 16 allowed values of NCS were chosen so
that the number of orthogonal preambles is as close as possible to what could be obtained if there
were no restrictions on the value of NCS..It is observed that the performance loss due to the
quantization is negligible.
Fig 3.5.4.1 - RACH Ncs set design

Figure shows the range of NCS values and their usage with the various preamble formats. NCS values
are designed for use in low-speed cells.
Cyclic Shift (NCS) Restriction for High-Speed Cells
The 64 RACH preambles assumes little or no frequency shifting due to Doppler spread, in the
presence of which ZC sequences lose their zero autocorrelation property. In the presence of a
frequency offset f , PRACH ZC sequence is distorted. A similar expression can be written for the
opposite frequency offset.
As can be observed, frequency offsets as large as one PRACH subcarrier (f =_fRA =
1/TSEQ = 1.25 kHz) result in cyclic shifts on the ZC sequence. This frequency offset f can be due
to the accumulated frequency uncertainties at both UE transmitter and eNodeB receiver, fLO, and the
Doppler shift resulting from the UE motion in a Line of Sight (LOS) radio propagation. The impact of
the cyclic shift distortion on the received Power Delay Profile (PDP) is, it creates false alarm peaks
whose relative amplitude to the correct peak depends on the frequency offset. The solution adopted in
LTE to address this issue is referred to as cyclic shift restriction and consists of masking some
cyclic shift positions in the ZC root sequence. This makes it possible to retain an acceptable false
alarm rate, while also combining the PDPs of the three uncertainty windows, thus also maintaining a
high detection performance even for very high-speed UEs.
It should be noted that at |f| = _fRA, the preamble peak completely disappears at the desired
location. However, the false image peak begins to appear even with |f |<_fRA. Another impact of the
side peaks is that they restrict the possible cyclic shift range so as to prevent from side peaks from
falling into the cyclic shift region.
We use C1 and C+1 to denote the two wrong cyclic shift windows arising from the frequency offset,
while C0 denotes the correct cyclic shift window. The cyclic restriction rule must be such that the two
Page 105
wrong cyclic shift windows C1 and C+1 of a cyclicly-shifted ZC sequence overlap none of C0, C1 or
C+1 of other cyclicly-shifted ZC sequences, nor the correct cyclic shift window C0 of the same cycliclyshifted ZC sequence, nor each other. Finally, the restricted set of cyclic shifts is obtained such that
the minimum difference between two cyclic shifts is still NCS but the cyclic shifts are not necessarily
multiples of NCS.
It is interesting to check the speed limit beyond which it is worth considering a cell to be a high-speed
cell. This is done by assessing the performance degradation of the PRACH at the system-level as a
function of the UE speed when no cyclic shift restriction is applied. A preamble detection is
considered to be correct if the timing estimation is within 2 s. A target Ep/N0 of 18 dB is used for the
first preamble transmission, with a power ramping step of 1 dB for subsequent retransmissions. The
cell radius is random between 0.5 and 12 km and a 2 GHz carrier frequency with frequency errors
within 0.05 ppm. The access failure rate is the measure of the number of times a UE unsuccessfully
re-tries access attempts (up to a maximum of three retransmissions), weighted by the total number of
new access attempts.
Fig 3.5.4.2 PCI and RootSequenceIndex Allocation for Neighbors

It is observed that under fading conditions,the RACH failure rates experience some degradation with
the UE speed (which translates into Doppler spread), but remains within acceptable performance
even at 350 km/h. For the AWGN channel (where the UE speed translates into Doppler shift) the
RACH failure rate stays below 102 up to UE speeds in the range 150 to 200 km/h. However, at 250
km/h and above, the throughput collapses.Without the cyclic shift restrictions the upper bound for
useful performance is around 150200 km/h.
Cyclic Shift Configuration for High-Speed Cells
The cyclic shift dimensioning for high-speed cells in general maximizes the sequence reuse when
group quantization is applied to cyclic shift values. However, for high-speed cells, the cyclic shift
restriction needs to be considered when deriving the sequence reuse factor with a specific cyclic shift
value. Note that there is no extra signalling cost to support an additional set of cyclic shift
configurations for high-speed cells since the one signalling bit which indicates a high-speed cell
configuration serves this purpose.
Page 106
Fig 3.5.4.3 Ncs index with Cell Radius derivation

The number of available preambles assumes no cyclic shift restriction at all, as in low-speed cells.
With the cyclic shift restriction above, the largest usable high-speed cyclic shift value among all root
sequences is 279.
Since for small NCS values the sequence usage is not so tight with a generally high sequence reuse
factor, a way to simplify design, while still achieving a high reuse factor, is to reuse the small NCS
values for normal cells. NCS values up to 46 are from the normal cyclic shift values, corresponding to
a cell radius up to 5.8 km. At the high end, the value of 237 rather than 242 is chosen to support a
minimum of two high-speed cells when all the 838 sequences are used. The maximum supportable
high-speed cell radius is approximately 33 km, providing sufficient coverage for preamble formats 0
and 2.
Sequence Ordering
A UE using the contention-based random access procedure needs to know which sequences are
available to select from. The full set of 64 sequences may require the use of several ZC root
sequences, the identity of each of which must be broadcast in the cell. Given the existence of 838
root sequences, signalling each individual sequence index requires 10 bits per root sequence, which
can lead to a large signalling overhead. Therefore in LTE the signalling is streamlined by broadcasting
only the index of the first root sequence in a cell, and the UE derives the other preamble signatures
from it given a predefined ordering of all the sequences.
Two factors are taken into account for the root sequence ordering, namely the CM of the sequence,
and maximum supportable cell size for high-speed cells (or equivalently the maximum supported
cyclic shift). Since CM has a direct impact on cell coverage, the first step in ordering the root
sequences is to divide the 838 sequences into a low CM group and a high CM group, using the CM of
QPSK (1.2 dB) as a threshold. The low CM group would be used first in sequence planning (and also
for high-speed cells) since it is more favourable for coverage.
Then, within each CM group, the root sequences are classified into subgroups based on their
maximum supportable cell radius, to facilitate sequence planning including high-speed cells.
Specifically, a sequence subgroup g is the set of all root sequences with their maximum allowed cyclic
shifts (Smax) lying between two consecutive high-speed NCS values according to
NCS(g) Smax <NCS(g + 1), for g = 0, 1, . . . , G 2, and
Smax NCS(G 1)
for G cyclic shift values, with the set of NCS values being those for high-speed cells. Sequences in
each subgroup are ordered according to their CM values.
Ordering arrangement reflects a continuous CM transition across subgroups and groups, which
ensures that consecutive sequences always have close CM values when allocated to a cell. Thus,
consistent cell coverage and preamble detection can be achieved in one cell. The LTE specifications
define the mapping from root sequence index u to a reordered index. The first 16 subgroups are from
the low CM group, and the last 16 from the high CM group.
Page 107
Sequences with Smax less than 15 cannot be used by any high-speed cells, but they can be used by
any normal cells which require no more than 24 root sequences from this group for a total of 64
preambles. Ordering of the physical ZC root sequence indices is pairwise, since root sequence
indices u and NZC u have the same CM and Smax values.
2.6.5 PRACH Implementation

UE Transmitter
The PRACH preamble can be generated at the system sampling rate, by means of a large IDFT. DFT
is optional as the sequence can be mapped directly in the frequency domain at the IDFT input. The
cyclic shift can be implemented either in the time domain after the IDFT, or in the frequency domain
before the IDFT through a phase shift. For all possible system sampling rates, both CP and sequence
durations correspond to an integer number of samples. The method leads to large IDFT sizes (up to
24 576 for a 20 MHz spectrum allocation), which are cumbersome to implement in practice.
Therefore, another option for generating the preamble consists of using a smaller IDFT, actually an
IFFT, and shifting the preamble to the required frequency location through time-domain up-sampling
and filtering (hybrid frequency/time-domain generation). Given that the preamble sequence length is
839, the smallest IFFT size that can be used is 1024, resulting in a sampling frequency fIFFT = 1.28
Msps. Both the CP and sequence durations have been designed to provide an integer number of
samples at this sampling rate. The CP can be inserted before the upsampling and time-domain
frequency shift, so as to minimize the intermediate storage requirements.
eNodeB PRACH Receiver

Front-End
In the same way as for the preamble transmitter, a choice can be made for the PRACH receiver at the
eNodeB between full frequency-domain and hybrid time/frequency domain approaches. The common
parts to both approaches are the CP removal, which always occurs at the front-end at the system
sampling rate fs, the PDP computation and signature detection. The approaches differ only in the
computation of the frequency tones carrying the PRACH signal(s).
The full frequency-domain method computes, from the 800 s worth of received input samples during
the observation interval, the full range of frequency tones used for UL transmission given the system
bandwidth. As a result, the PRACH tones are directly extracted from the set of UL tones, which does
not require any frequency shift or time domain filtering but involves a large DFT computation. Note
that even though NDFT = n 2m, thus allowing fast and efficient DFT computation algorithms inherited
from the building-block
construction approach, the DFT computation cannot start until the complete sequence is stored in
memory, which increases delay.
On the other hand, the hybrid time-frequency domain method first extracts the relevant PRACH signal
through a time-domain frequency shift with down-sampling and anti-aliasing filtering. There follows a
small-size DFT (preferably an FFT), computing the set of frequency tones centered on the PRACH
tones, which can then be extracted. The down-sampling ratio and corresponding anti-aliasing filter are
chosen to deliver a number of PRACH time samples suitable for an FFT or simple DFT computation
at a sampling rate which is an integer fraction of the system sampling rate. Unlike the full frequencydomain approach, the hybrid time/frequency-domain computation can start as soon as the first
samples have been received, which helps to reduce latency.
Power Delay Profile Computation
The LTE PRACH receiver can benefit from the PRACH format and Constant Amplitude Zero
AutoCorrelation (CAZAC) properties by computing the PRACH power delay profile through a
frequency-domain periodic correlation.
The zero padding aims at providing the desired oversampling factor and/or adjusting the resulting
number of samples to a convenient IFFT size. Note that for high-speed cells, additional non-coherent
combining over three timing uncertainty windows can be performed for each receive antenna.
In addition, the pairwise sequence indexing allows further efficient paired matched filtering. The
IDFTs of the joint computation of the frequency domain matched filters provide the periodic
correlations of time domain ZC sequences xu(n) and xNZCu(n), the latter being shifted by one
sequence sample.
Signature Detection
Page 108
The fact that different PRACH signatures are generated from cyclic shifts of a common root sequence
means that the frequency-domain computation of the PDP of a root sequence provides in one shot
the concatenated PDPs of all signatures derived from the same root sequence. Therefore, the
signature detection process consists of searching, within each ZCZ defined by each cyclic shift, the
PDP peaks above a detection threshold over a search window corresponding to the cell size.
Detection threshold setting.
The target false alarm probability pfa(Tdet) drives the setting of the detection threshold Tdet. Under the
assumption that the L samples in the uncertainty window are uncorrelated Gaussian noise with
variance 2n in the absence of preamble transmission, the complex sample sequence zma ( )
received from antenna a (delayed to reflect a targeted time offset of the search window, and
despread over a coherent accumulation length (in samples) Nca against the reference code
sequence) is a complex Gaussian random variable with variance 2n,ca = Nca2n . The relative
detection threshold can be precomputed and stored.
Noise floor estimation.
In a real system implementation, the number of additions can be made a power of two by repeating
some additions if needed. The initial absolute threshold Tdet_ini is computed using an initial noise floor
estimated by averaging across all search window samples.
Collision detection.
In any cell, the eNodeB can be made aware of the maximum expected delay spread. As a result,
whenever the cell size is more than twice the distance corresponding to the maximum delay spread,
the eNodeB may in some circumstances be able to differentiate the PRACH transmissions of two UEs
if they appear distinctly apart in the PDP. Collision detection is never possible, while the lower PDP
represents a larger cell where it may sometimes be possible to detect two distinct preambles within
the same ZCZ. If an eNodeB detects a collision, it would not send any random access response, and
the colliding UEs would each randomly reselect their signatures and retransmit.
Timing Estimation
The primary role of the PRACH preamble is to enable the eNodeB to estimate a UEs transmission
timing. One can observe that the timing of 95% of UEs can be estimated to within 0.5 s, and more
than 98% within 1 s. No collision detection algorithm is implemented here. The IFFT size is 2048 and
the system sampling rate 7.68 MHz, giving an oversampling rate of 2.44.
Channel Quality Estimation
For each detected signature, the relative frequency-domain channel quality of the transmitting UE can
be estimated from the received preamble. This allows the eNodeB to schedule the L2/L3 message
(message 3) in a frequency-selective manner within the PRACH bandwidth.
The BLER performance of the L2/L3 message of the RACH procedure when frequency-selectively
scheduled or randomly scheduled, assuming a typical 10 ms delay between the PRACH preamble
and the L2/L3 message. A Least Squares (LS) filter is used for the frequency-domain interpolation,
and a single RB is assumed for the size of the L2/L3 message. It can be seen that the performance of
a frequency-selectively scheduled L2/L3 message at 10% BLER can be more than 2 dB better than
blind scheduling at 3 km/h, and 0.5 dB at 10 km/h.
2.6.6 Time Division Duplex (TDD) PRACH

One design principle of LTE is to maximize the commonality of FDD and TDD transmission modes.
With this in mind, the random access preamble formats 0 to 3 are supported in both FDD and TDD
operation. In addition, a short preamble format,format 4, is supported for TDD operation. Format 4 is
designed to fit into the short uplink special field known as UpPTS for small cells.
Preamble Format 4
For preamble format 4, a ZC sequence of length 139 is used. The preamble starts 157 s before the
end of the UpPTS field at the UE. Unlike preamble formats 0 to 3, a restricted preamble set for highspeed cells is not necessary for preamble format 4, which uses a 7.5 kHz subcarrier spacing. With a
random access duration of two OFDM symbols (157 s), the preamble format 4 is mainly used for
small cells with a cell radius less than 1.5 km, and where cyclic shift restrictions for high UE velocities
Page 109
are not needed. Therefore, considering that Layer 2 always sees 64 preambles and a sequence
length of 139, a smaller set of cyclic shift configurations can be used.
Unlike for preamble formats 0 to 3, the root ZC sequence index for preamble format 4 follows the
natural pairwise ordering of the physical ZC sequences, with no special restrictions related to the CM
or high-speed scenarios.
Cyclic shift configuration for preamble format 4.
NCS NCS value
Required number of ZC root sequences per cell
-------- -------------- -- -------------------------------------------------------------0
2
1
1
4
2
2
6
3
3
8
4
4
10
5
5
12
6
6
15
8
It can be seen how the PRACH preamble addresses the high performance targets of LTE, such as
high user density, very large cells, very high speed, low latency and a plurality of use cases, while
fitting with minimum overhead within the uplink SC-FDMA transmission scheme. Many of these
aspects benefit from the choice of ZC sequences for the PRACH preamble sequences in place of the
pseudo-noise sequences used in earlier systems. The properties of these sequences enable
substantial numbers of orthogonal preambles to be transmitted simultaneously.
Considerable flexibility exists in the selection of the PRACH slot formats and cyclic shifts of the ZC
sequences to enable the LTE PRACH to be dimensioned appropriately for different cell radii and
loadings. Some options for the implementation are available, by which the complexity of the PRACH
transmitter and receiver can be minimized without sacrificing the performance.
2.6.7 Uplink Timing Control

A key feature of the uplink transmission scheme in LTE is orthogonal multiple-access in time and
frequency between the different UEs.
Propagation delays in uplink and downlink are absorbed at the NodeB, by means of reducing the time
spent measuring the Signal-to-Interference Ratio (SIR) to derive the next power control command.
Uplink orthogonality is maintained by ensuring that the transmissions from different UEs in a cell are
time-aligned at the receiver of the eNodeB. This avoids intracell interference occurring, both between
UEs assigned to transmit in consecutive subframes and between UEs transmitting on adjacent
subcarriers. Time alignment of the uplink transmissions is achieved by applying a timing advance at
the UE transmitter, relative to the received downlink timing. The main role of this is to counteract
differing propagation delays between different UEs. A similar approach is used in GSM.
2.6.8 Timing Advance Procedure

Initial Timing Advance
After UE has first synchronized to the DL PSS and SSS, the initial timing advance is set by RA
procedure. This involves UE transmitting a RACH and eNodeB can estimate UL timing and respond
with an 11-bit initial timing advance (TA) command within Random Access Response (RAR) with a
granularity of 0.52 s from 0 up to a maximum of 0.67 ms, corresponding to a cell radius of 100 km.
Larger cell radii would in theory be possible, but a cell range of 100 km is sufficient for practical
scenarios, and is far beyond what could be achieved with the early versions of GSM in which the
range of the timing advance restricted the cell range to about 35 km.
1. For RA response, 11-bit TA command, TA, indicates NTA of TA = 0, 1, 2, ..., 1282.
1. where amount of time alignment is given by NTA = TA 16.
2. In other cases, 6-bit TA command, TA, indicates NTA, NTA,old, to the new NTA, NTA,new,
by index of TA = 0, 1, 2,..., 63,
1. where NTA,new = NTA,old + (TA 31)16.
2. Adjustment of NTA value indicates advancing(+) or delaying(-) the UL Tx
timing.
Page 110
The granularity of 0.52 s, enables an accuracy, well within CP length (smallest CP is 4.7 s) and
also finer than the a cyclic shift of UL RS. It is observed that timing misalignment of up to at least 1 s
does not cause significant degradation in system performance due to increased interference. Thus
granularity of 0.52 s is sufficiently fine.
Timing Advance Updates
After the timing advance has first been set for each UE, it will then need to be updated from time to
time to counteract changes in UL signals timing at eNB. Such changes may arise from:
1. Fast movement of UE, causing propagation delay to change at a rate relative to eNodeB; at
500 km/h, the round-trip propagation delay would change by a maximum of 0.93 s/s.
2. Abrupt changes in propagation delay due to existing propagation paths disappearing and new
ones coming into play; like in dense urban as UEs move around the corners of buildings.
3. Oscillator drift in UE with frequency errors accumulated over time result in timing errors; the
frequency accuracy in UE is required to be better than 0.1 ppm, maximum accumulated
timing error of 0.1 s/s.
4. Doppler shift arising from the movement of the UE, results in an additional frequency offset of
the UL signals. TA to counteract these effects are performed by a closed loop mechanism
whereby eNodeB measures received UL timing and issues TA commands to instruct the UE
to adjust its timing relative to its previous transmission timing.
In deriving the timing advance update commands, the eNodeB may measure any uplink signal which
is useful. This may include SRSs, CQI, ACK/NACK for DL data, or UL data themselves. Highly
accurate timing estimation has to be traded off against UL overhead from such signals. Cell-edge UEs
are power limited and therefore also bandwidth-limited for a given UL SINR;
A TA command received at the UE is applied at the beginning of UL subframe which begins 45 ms
later. For TDD or half-duplex FDD, the new timing would take effect at the start of the first uplink
transmission after this point. In case of increase in TA relative to previous, first part of subframe in
which new timing is applied is skipped.
The TA commands are generated at MAC layer in eNodeB and transmitted to UE as MAC packets
which may be multiplexed with data on PDSCH. Update TA commands have a granularity of 0.52 s.
The range of update commands is 16 s, allowing a step change in timing equivalent to the length of
extended CP. They would typically not be sent more frequently than about 2 Hz, not too frequent.
The eNodeB must balance the overhead of sending regular timing update commands to all the UEs in
the cell. The eNodeB therefore configures a timer for each UE, which UE restarts each time a TA is
received; if UE does not receive another TA before timer expires, it declares uplink to have lost
synchronization. In such a case, UE is not permitted UL transmission without first transmitting a
random access preamble to reinitialize the uplink timing.
2.6.9 Power Control

Overview
Uplink power control balances the need for sufficient transmitted energy per bit to achieve required
QoS, against needs to minimize interference and to maximize battery life of UE. Uplink power control
has to adapt to propagation channel, path loss, shadowing and fast fading, as well as interference
from other users both intra and inter cells.
In WCDMA, UL is non-orthogonal and it has to manage intra-cell UE interference sharing the
same time-frequency resources. To increase uplink data rate for a given user in WCDMA, UE
need to reduce the spreading factor and increase the power accordingly. By contrast, in LTE,
UL is orthogonal, and intra-cell interference is less critical. To vary UL data rate, vary the
bandwidth and MCS, while the Tx power per unit bandwidth (i.e. Power Spectral Density
(PSD)) could remain constant for a given MCS.
In WCDMA, power control was with continuous Tx for circuit-switched services and is periodic
with a loop delay of 0.67 ms and a normal power step of 1 dB. In LTE 1 ms intervals fast
data UE switching is required. it allows for larger power steps (non periodic), with a
minimum loop delay of 5 ms.
LTE employs a combination of open-loop and closed-loop control. This requires less feedback than a
purely closed-loop and is used to compensate for cases when the UEs own estimate of required
power setting is not satisfactory.
UE sets a coarse operating point for Tx PSD by open-loop, based on path-loss. Faster adaptation is
applied around open-loop operating point by closed loop power control. This can control interference
Page 111
and fine-tune the power setting to suit the channel conditions (including fast fading). Due to
orthogonal nature of UL, power control does not need to operate fater than a few hundred Hertz.
Fastest and most frequent adaptation is by UL scheduling grants, which vary bandwidth (and total Tx
power) and MCS.
Detailed Power Control Behaviour

Power control formulae are specified for PUSCH, PUCCH and SRSs, all with same basic principles.
(1) a basic open-loop operating point derived from parameters signalled by eNodeB, and (b) a
dynamic offset subframe to subframe:
Power/RB = basic open-loop operating point + dynamic offset
Basic Open-Loop Operating Point
The basic open-loop operating point depends on inter-cell interference and cell load. Two main
components are:
1. a semi-static base level, P0, comprised of a common power level for all UEs in the cell
(measured in dBm) and a UE-specific offset;
2. an open-loop path-loss compensation component.
PUSCH base levels are dynamically scheduled (by PDCCH) by RRC signalling, allowing different
BLER (BLock Error Rate) operating points. For persistently-scheduled transmissions like VoIP, it
avoids PDCCH signalling overhead.
UE-specific offset of base level P0 enables eNB to correct systematic offsets needed for errors in
path-loss estimation or in absolute output power setting.
Path-loss compensation is based on UEs estimate of DL path-loss, derived from RSRP and known
Tx power of DL RSs, broadcast by eNodeB. To obtain path-loss, UE should filter DL path-loss with a
suitable time-window (100-500ms) to remove effect of fast fading.
For PUSCH and SRS, the degree to which UL PSD is adapted to compensate path-loss () is set by
eNB (no compensation to full compensation).
The base level P0 and path-loss compensation () together are coordinated with UE measurements.
At one extreme, eNodeB could configure base level to lowest level (126 dBm) and rely entirely on
UEs path-loss measurement to raise power towards the cell edge, while alternatively eNodeB can set
the base level to a higher value, in conjunction with only partial path-loss compensation. Base level P0
for PUSCH (126 dBm to +23 dBm per RB) is designed to cover the full range of target SINR for
different path-loss compensation, bandwidths and interference. For example:
highest P0, +23 dBm, corresponds to maximum Tx power of LTE UE, and used if path-loss
compensation was not used at all.
Lowest P0 for PUSCH, 126 dBm, is relevant when full pathloss compensation is used and
UL Tx and Rx conditions are optimal. Taking a single RB transmission, with target SINR at
eNB of 5 dB (around lowest useful SINR), interference-free reception and a 0 dB noise
figure for eNB, then required P0 is thermal noise level in one RB minus 5 dB, which gives P0
=126 dBm.
Maximum path-loss compensation (P0 or path-loss ) depends on the required SINR and BW. For
SINR from 5 dB to +30 dB, interference rise above thermal noise from 0 dB to +30 dB, and BW from
one RB to 110 RBs (19.8 MHz).
Fractional path-loss compensation, is a trade off of fairness of UL scheduling against total cell
capacity. Full path-loss compensation maximizes fairness for cell-edge UEs. However, only partial
path-loss compensation can increase the total Network capacity in UL, as less resources are spent
ensuring cell-edge UEs transmission and less inter-cell interference is caused. Path-loss
compensation factors around 0.70.8 typically give a close-to-maximal UL system capacity without
significant degradation to the cell-edge achievable data rate.
UEs located near the edge of a cell, may disrupt UL transmissions in neighbouring cells. A frequencydependent overload indicator may be signalled directly between eNodeBs to warn a neighbouring
eNodeB of high uplink interference levels in specific RBs. In response, neighbouring eNodeB may
reduce the permitted energy per RB of the UEs scheduled in the corresponding RBs in its cell(s). It is
also possible to cooperate to avoid scheduling cell-edge UEs in neighbouring cells to transmit in the
same RB.
In summary, Basic operating point = P0 + PL
For low-rate PUCCH (with ACK/NACK and CQI), path-loss compensation is handled separately from
PUSCH, as PUCCH from different users are code-division-multiplexed. To provide good control of
interference between different users, and to maximize number of users simultaneously on PUCCH,
Page 112
full path-loss compensation is always used. A different base level P0 is also provided for PUCCH
compared to those used for the PUSCH.
Dynamic Offset
The dynamic offset (Per RB) = MCS dependent component + TPC commands.
MCS-dependent component - The MCS-dependent component (dTF,Transport Format) allows
additional transmitted power per RB for that data rate. Normalized data rate as Shannons theorem:
RN = log2(1 + SNR)* 1/k = bits per RE = BPRE and
SNR = Signal-to-Noise Ratio, k= 1.25.
MCS-dependent component is like a power control command, as the MCS is controlled by eNodeB
scheduler. MCS-dependent component can also be like frequency dependent power control when
scheduling a low-rate MCS in a particular part of the band. eNodeB can dictate a low transmission
power in those RBs.
Uplink RBs allocated to a UE in a subframe may not be matched to the desired data rate and SIR.
Enable transmit power to be reduced if data transmitted is less than the rate supported by the radio
channel in a single RB. The MCS-dependent component for the PUSCH can be set to zero if fast
Adaptive Modulation and Coding (AMC) is used instead.
PUCCH bandwidth for a UE does not vary (although it may be ranging from a 1 bit SR/ACK/NACK, to
22 bits combined ACK/NACK and CQI together). Magnitude of power offset for each control can be
adjusted semi-statically by eNodeB to set a suitable error-rate operating point.
UE-specific power control commands. Other dynamic offset is UE-specific TPC commandsaccumulative (for PUSCH, PUCCH and SRS) and absolute (for PUSCH only). For PUSCH, switch
between these two modes is done by RRC for each UE.
1. Accumulative - each TPC signals a power step relative to previous level. This is default and
is suited when successive subframes received as closed loop. In LTE, two sets of power step
values are provided: either {1, +1} dB or {1, 0, +1, +3} dB configured by TPC command and
RRC configuration. Maximum power step size= +3/1 dB, adjustable upto maximum and
minimum power limits according to the UE power class. A 0 dB step size means transmit
power to be kept constant if needed.
2. Absolute - TPC command is independent of sequence of TPC commands received
previously; depends only on the most recently-received absolute TPC command, a power
offset relative to the semi-static operating point. Absolute TPC commands offset set is {4,
1, +1, +4} dB.
The absolute power control mode can only control power within 4 dB from operating point, but a
relatively power step can be triggered by a single command (up to 8 dB), suited to scenarios of
intermittent transmission. Absolute TPC command adjusts to a suitable level in a single step.
Total Transmit Power Setting
Finally, for PUSCH and SRS, total transmit power in each subframe is scaled up linearly according to
RBs scheduled from UE in the subframe. Thus the overall power control equation is as follows:
UE Tx power = P0 + *PL(basic open-loop operating point)
+ dTF + f (dTPC)(dynamic offset)
+ 10 log10 M (bandwidth factor)
where dTPC TPC command, f () accumulation and M =# RBs.
This overall power control formula allows UEs transmit power is allowed to be, typically 50 dBm to
+23 dBm (0.2 W).
Transmission of TPC Commands
TPC commands are sent to UE in PDCCH messages. UE is required to check for TPC command in
every subframe unless DRX is configured. However, TPC commands are not necessarily periodic.
TPC commands may be transmitted in UL scheduling assignment messages for each UE, resulting in
information for an uplink transmission (RB, MCS, and power) in a single message.
Individual accumulative TPC commands for multiple UEs can be jointly coded into a special PDCCH
message Formats 3 and 3A; furthermore, for PUCCH only, TPC commands can be sent in DL
assignment messages on PDCCH. These methods track changes in channel conditions even when
UE is not scheduled for UL, and works as alternative to absolute TPC. No jointly-coded TPC allowed
on PDCCH in absolute power control mode.
Page 113
PDCCH signalling with TPC commands are protected by a CRC and are reliable. The eNodeB may
use many techniques to determine transmit power. One method will be received SIR measurements
of SRS and UL RSs and BLER of uplink data. Also take into account interference coordination with
neighbouring cells, overload indicator (interference from a UE is causing a problem).
UE Power Headroom Reporting

To assist eNB to schedule UL RB to different UEs, UEs report its available power headroom - how
much more UL bandwidth per subframe a UE is capable of using. If more headroom is there, it may
allocate more power, instead of more RBs to save resources.
Power headroom range is from +40 to 23 dB. Negative range signals the extent to which it has
received UL grant requiring more power than available. This would enable eNodeB to reduce the RB
size of a subsequent grant, thus freeing up transmission resources to allocate to other UEs.
Headroom report is a prediction (not measurement), relies on reasonably accurate calibration of the
UEs power amplifier output.
Power headroom report criteria:
1. Significant change in path loss since last report;
2. Configured time has elapsed since last report;
3. Configured number of closed-loop TPC commands are implemented by UE.
The eNodeB can configure these parameters to control these triggers depending on loading and
scheduling algorithm.
PUSCH Power Calculation
Transmit power will be calculated at each TTI (subframe i) for each carrier c as follows:
where, i = subframe, c=cell-id, C= #codeblocks, Kr= Size of r-th codeblock

Where, O(cqi)= No. Of CQI/PMI bits(incl CRC), N(re)= #RE =M(sc)*N(symb).
st
nd
Mark that, if no simultaneous PUCCH is there, then, 1 and 2 options are same.
P(cmax,i) = Default Configured UE Tx Power in subframe i;
P(pucch,i) = Power allocated for PUCCH in i.
M(pusch.i) = Bandwidth of PUSCH #RB for subframe i in cell c.
P(o_pusch, j) = P(o-nominal_pusch , j) + P(o_ue_pusch, j)) for j=0,1
o j=0 for Tx in semi-persistent grant
o J=1 for Tx in Dynamic Scheduled grant
o J=2 for Tx in RA grant. And P(o_ue_pusch, 2)=0.
Alpha(j) = {0, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9 or 1}, 3bit factor for path loss
PL = DLPathLoss calculated by UE in dB = Ref Sig
(PathLossReferenceLinking)
Pwr
RSRP.
Page 114
BPRE.Ks
Delta TF(i)= 10 log( (2

-1). B(offset,pusch) )
o Ks=0 for Tx mode=2 and Ks=1.25 for rest. Ks is derived from. Delta_MCS_enabled
o BPRE= O(qci)/N(re) for control data via PUSCH w/o UL-SCH, OR
o BPRE= sum(over r=0..C-1)(Kr/N(re)) for all other cases.
o B(offset,pusch)=B(offset,cqi) for control data via PUSCH w/o UL-SCH and 1 for
others.
Correction delta(pusch, c). is referred as TPC command included in PDCCH with
o DCI format 0/4 for serving cell or jointly
o DCI format 3/3A, CRC parity bits scrambled with TPC-PUSCH-RNTI.
o Current PUSCH power control adjustment is given by F(i).
f(i) = f(i-1) + delta(pusch, (i-T)). Or
No f(i-1) is added if no accumulation indicated.
delta(pusch, (i-T)) is the TPC received T subframe ago.
PL Pathloss = referenceSignalPower higher layer filtered RSRP,
o Ks is derived from deltaMCS-Enabled . For Tx mode 2, Ks=0.
First Original value of P(o_ue_pusch) is set as follows:
o Initially f(0) = 0 or whatever is received from P-cell as a change value.
o For P-Cell: f(0)= delta_P(rampup) + delta(msg2) from the RA response process.
Delta(msg2) = TPC command indicated in RA response, and
Delta_P(rampup) = total power ramp-up from first to the last preamble
PUCCH Power Control
The formula for PUCCH would be:
P_PUCCH(i) = min{P_CMAX, P_0_PUCCH + PL + h(n_CQI, n_HARQ) + Delta_F_PUCCH(F) + g(i)}
i : Subframe Number
j : This can be 0 or 1
P_O_PUCCH: P_O_NOMINAL_PUCCH + P_O_UE_PUCCH, where P_O_NOMINAL_PUCCH and
P_O_UE_PUCCH came from higher layer(RRC Connection Setup or RRC ConnectionReconfig).
Summary of Uplink Power Control Strategies

Each deployment will select a appropriate mode of power control. One mode may be semi-static
operating point (via P0 and PL compensation factor ) to achieve required SINR at eNodeB for
required QoS for each UE.
Further interference management and rate adaptation is managed by RB scheduling giving degrees
of freedom for power management. It may change MCS to set different BLER operating points for
different HARQ processes.
Finally, it may use MCS-dependent offsets and the closed-loop TPC commands. This gives full
flexibility of power control.
Page 115
2.6. Miscellaneous
2.6.1 Evaluation LTE Physical Layer Questions
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
What is the content of the cyclic prefix?

How wide is the MBMS subcarrier wide?
How many symbols are there in 1 millisecond in Random Access subcarrier?
What does CSI consist of?
What information is derived out of PSS/SSS signals?
Where are PSS/SSS mapped in Physical resources?
Where is PBCH mapped? How many RB's do they occupy?
Describe briefly how, where and how many padding 0 bits are inserted in code
segmentation.
Can control information be sent on PUSCH without UL-SCH data?
What is the CRC length, total number of segments, Encoder type and Modulation for
PBCH?
In which CSI transmission case PUCCH is used?
CQI is 4 bits. What does 0, 1 and 15 indicate respectively.
How many DMRS are there in a RB in a PUCCH of type 1/1a/1b and of type 2/2a/2b?
Explain why a mobile with speed more than 500Km/Hour will fail to remain connected
with LTE?
What is a preamble format? How many types are there? Explain when which type of
format is suitable?
What is the length of preamble sequence in bits and how many RB's are required to
send?
what are the values transmitted as CSI?
What is SRS and where is it physically mapped? Explain.
What is sent in DCI format 0?
How many symbols and REGs does PCFICH occupy?
how many REGs does one PHICH occupy and where in the DL Physical Resource?
How many mobiles PHICH can be sent in the same symbol and how are they
multiplexed?
How is PHICH and CFICH allocated in different cells so that they avoid interference.
Show it pictorially to explain.
In case of 4x4 MIMO, how many reference signals are there in an RB?
What is the %tage occupied by Ref signals in 4x4 MIMO case?
In FDD, the UL and DL frame timing is slightly skewed in time and who is earlier by how
much?
Which are the Physical signals sent in UCI?
Which are the signals sent in DCI?
What is RV (redundancy version)?
What is TTI Bundling.
What is PMI and
What is RI?
What are the valid CRC length in bits?
For a buffer of length 6140 from MAC (TTI), what is the number of segments? Why?
Which channel case Turbo coding is used?
which channel case BitRepeatationCoding is used?
What is the smallest and biggest segment size for Encoding?
For Turbo Encoder - If block size is 80, what is the number of output bits.
For Convlutional Encoder- If block size is 80, what is the number of output bits?
How many trellis bits are additional terminated with and with which streams?
Page 116
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
What is Trellis Termination for Turbo Encoder?

Total Number of bits G and mapped onto G'=G/(L*M) symbols. What is L and M?
For Turbo coding rate matching, data block is to be spread in how many columns?
What are the control information multiplexed with data of PUSCH?
what are the control information interleaved together with PUSCH?
In DL, How many RS symbols are there in a normal RB in case of 2x2 MIMO?
What is EPRE?
What is TPC and TPC accumulation?
What is power headroom and
Which symbol locations are occupied for DMRS in a PUCCH of type 1 and type 2? (start
counting from 0).
How many Symbols are used for PUSCH in a SRS subframe per subcarrier, including
multiplexed CQI.
For DL-SCH, PMCH and PCCH, what is CRC length,number of segments, Encoder type
and modulation?
What is the CRC length for CFICH and PHICH respectively?
In which situation 4 control signals are used for PDCCH and why?
What kind of encoder is used for CFICH and PHICH?
For RAR, how many bits is TA?
If the TA is 25 in the RAR, what is the TA adjustment required in microseconds?
How many bits is TA adjustment for non-RAR Timing Advance?
If the TA value is 25(decimal), how much milli/ microsecond timing is to be adjusted?
If TA=25, Is the timing to be increased or decreased?
Normally, when should the TA adjustment should take place, if you receive it in subframe
x?
If you receice a Scell Activation or De-activation in subframe x, when should that be
effective?
In EPRE calculation what is the relationship with Cyclic Prefix?
What are the main factors in calculation of Power for PUSCH? List them.
What are the 3 parameters, based on which scrambling code is generated?
What is semipersistent signalling?
What is DRX and DRX active period.
What is TTI bundling?
What is HARQ RTT Timer?
What is the max TPC accumulation(dbm) value in both +ve and -ve direction?
What is Power scaling factor?
Cell Parameter P(B) is denoted by (B)/(A). What are these 3 parameters w.r.t. symbol
EPRE?
What is the maximum P(B)?
List the parameters UE must have before Random Access?
What are the factors which determine the initial RA power?
After getting RAR, What do you send to eNB?
First L3 message after RAR is sentafter how long delay from RAR?
If you do not receive RAR within RAR window, after/before what delay would you
retransmit?
What is the significance of "UL delay" and "hopping flag" in RAR?
What is the max number of DL HARQ processes for FDD and TDD in UL/DL both?
How many Maximum bits are there for PUCCH of format 1/1a/1b and 2/2a/2b each and
they are spread in how many symbols?
Explain difference between Resource allocation type 0 and 1?
What is RBG(Resource Block Group) and size of each RBG?
What are the parameters defining the TBS size?
Page 117
85 What is the minimum(with minimum modulation) and Maximum(with maximum

modulation) TBS (Transport Block Size) for one layer.
86 If for single layer TBS size is "X", will the TBS size be n.X for n layers? Why?
87 What are CSI elements?
88 What is Periodic and Aperiodic CSI?
89 Max how many MACs can be configured for TTI bundling?
90 In a group hopping, how many hopping patterns are there and how many sequence shiftpatterns?
91 For every subframe of PUSCH how is ack/nack sent?
92 for every PDSCH, when/how is the ack/nack sent?
93 List the types of RNTI (with their default values, if any).
94 What are parameters defining SRS?
95 What is the difference between type 1 and type 2 hopping?
96 What is difference between intra-subframe and inter-subframe hopping?
97 What are the 4 configurations of CCE?
98 How will you find number of PDCCH accomodatable in a subframe?
99 What is Common Search space?
100 What is UE-Specific Search Space?
101 How is PHICH assigned on the control region?
102 How is CFICH assigned in the control region?
103 How many PUSCH DMRS RE's are there in a RB in a subframe?
104 Where are DMRS mapped on the PUSCH RB in a subframe?
105 What is cyclic shift?
106 How is cyclic shift it used for PHICH or other control information among UE's?
107 What is the spreading factor used for PHICH after repeat encoding?
108 In PHICH what is the total number of symbols occupied?
109 DL Phy Layer transmission is Synchronous or asynchronous?
110 UL Phy Layer transmission is Synchrnous or Asynchronous?
111 What is semi-persistent scheduling?
112 How many bits are used to send Scheduling Request in PUCCH?
113 How is SR sent and multiplexed with other control information?
114 How many users PDCCH can be multiplexed for format 1 and format 2 respectively on
the same physical resource?
115 In rel-10 how many antenna max assumed for UE?
116 In rel-10 How many antenna max in eNB?
117 How many types of reference signals are there in Downlink? List them.
118 In case of 4 antennas with 4x4MIMO, maximum how many codewords and layers can be
configured?
Page 118
Page 119
Page 120

LTE Protocol Stack Physical Layer

Загружено:

Сведения о документе

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

LTE Protocol Stack Physical Layer

Загружено:

Авторское право:

Доступные форматы

LTE Development, Conformance Test, Optimization

Certification Course Amateur Level (3PCA-L1)

LTE Protocol Stack- 1

Who is this book for?

LTE Protocol Stack Development

Edition: July 2013

LTE Physical Layer- 3PCA-L1 Certification

Downlink Physical Layer ................................................................................................................. 7

OFDMA Principles ................................................................................................................... 7

OFDM - Orthogonal Multiplexing Principle ...................................................................... 7

Peak-to-Average Power Ratio and Sensitivity ................................................................ 9

Timing Offset and Cyclic Prefix Dimensioning ................................................................ 9

OFDMA Parameter Dimensioning .............................................................................. 10

Physical Layer Parameters for LTE .............................................................................. 10

Transmission Resource Structure ................................................................................. 11

Synchronization and Cell Search .......................................................................................... 15

Synchronization Sequences and Cell Search in LTE ................................................... 15

ZadoffChu Sequences ................................................................................................ 16

Primary Synchronization Signal (PSS) Sequences ...................................................... 17

PSS Generation ............................................................................................................ 17

Secondary Synchronization Signal (SSS) Sequences.................................................. 18

Cell Search Performance .............................................................................................. 19

Reference Signals and Channel Estimation ................................................................. 19

Design of Reference Signals in LTE ............................................................................. 19

Cell-Specific Reference Signals (CRS) ......................................................................... 20

UE-Specific Reference Signals(URS) ........................................................................... 21

RS-Aided Channel Modelling and Estimation ............................................................... 22

Frequency Domain Channel Estimation ....................................................................... 22

Time-Domain Channel Estimation ................................................................................ 22

Spatial Domain Channel Estimation(SD-MMSE) .......................................................... 23

Phy Data and Control Channels - DL .................................................................................... 25

Physical Broadcast Channel (PBCH) ............................................................................ 25

Physical Downlink Shared Channel (PDSCH) .............................................................. 26

Physical Multicast Channel (PMCH) ............................................................................. 27

Downlink Control Channels ........................................................................................... 27

Physical Control Format Indicator Channel (PCFICH).................................................. 28

Physical Downlink Control Channel (PDCCH) .............................................................. 29

PDCCH Candidate Selection ........................................................................................ 30

Formats for Downlink Control Information (DCI) ........................................................... 31

Physical Hybrid ARQ Indicator Channel (PHICH) ......................................................... 35

Resource Allocation ...................................................................................................... 36

DL Resource Allocation Rules ...................................................................................... 38

Resource Allocation Bitmap examples.......................................................................... 39

Uplink Grant .................................................................................................................. 40

LTE Physical Layer- 3PCA-L1 Certification

PDCCH Transmission and Blind Decoding ................................................................... 41

Enhanced PDCCH (EPDCCH) ...................................................................................... 41

EPDCCH assignment procedure .................................................................................................. 41

Physical Layer Processing - DL ............................................................................................ 47

Link Adaptation and Feedback Computation ................................................................ 48

CQI Feedback in LTE .................................................................................................... 48

Viterbi Algorithm (VA) (Example): ................................................................................. 49

LTE Contention-Free Interleaver................................................................................... 51

Coding for Control Channels in LTE ............................................................................. 54

General structure for downlink physical channels ......................................................... 54

Mapping to resource elements ...................................................................................... 56

Scheduling Process - Control Channel Viewpoint ........................................................ 45

MIMO Techniques ................................................................................................................. 59

Introduction to MIMO ..................................................................................................... 59

Single-User (SU-) MIMO Techniques ........................................................................... 60

Multi-User Techniques .................................................................................................. 63

MIMO Schemes in LTE ................................................................................................. 64

Single-User Schemes ................................................................................................... 64

Beamforming Schemes ................................................................................................. 65