Вы находитесь на странице: 1из 66

Frequency Hopping in Nokia BSS

© NOKIA FH_SHOW.PPT
FH in Nokia supplied networks

FH has not been utilised in 'Nokia' NW's due to:


• problems with mobiles
• Intelligent Overlay-Undelay providing the needed capacity boost

Now FH is more viable as the IUO/FH is coming and the FH support


in general is better in BSS.

Frequency

FH is currently used by several F1


operators, both in live networks and trial. F2
F3
F4

Time

© NOKIA FH_SHOW.PPT
What is Frequency Hopping?

Call is transmitted through several


Frequency frequencies in order to
• average the interference (interference diversity)
• minimise the impact of fading (frequency diversity)

F1

F2

F3

Time

© NOKIA FH_SHOW.PPT
Hopping principles

Frequency

RANDOM CYCLIC

F3

F2

F1
cycle

Time

RANDOM - better in cities CYCLIC - better in rural

© NOKIA FH_SHOW.PPT
The Benefits of Frequency Hopping?
1. Frequency diversity; Compensates the frequency
selective fading
Signal
Level

F1
F2
F3

MS Location Distance

Bursts sent on frequency F2 are degraded or lost, but the initial signal may
still be reconstructed from the bursts on frequencies F1 and F3.

© NOKIA FH_SHOW.PPT
Fading

long term fading


(log-normal fading) short term fading
(multipath fading)

Amplitude

Long term
fading

Short term
fading
Distance

© NOKIA FH_SHOW.PPT
The benefits of Frequency Hopping?
2. Interference diversity; the interference is averaged
over multiple frequencies

Interference Interference With hopping


No hopping
F1
F1
F3 F1

F1

F2 average
F2 F3 F2 F3
F2 F
3

MS_1 MS_2 MS_3 MS_1 MS_2 MS_3

© NOKIA FH_SHOW.PPT
Capacity gain through interference diversity

Excess C/I margin can be used for


traffic capacity increase through
smaller reuse factor
Level

Carrier

15dB(90%LP) 9dB
>>15dB
Max
Interference
Min Average
interference

No Hopping With Hopping With Hopping

Tighter reuse

© NOKIA FH_SHOW.PPT
FH Gain
Relative to number of hopping frequencies

8
7
6
Gain [dB]

5 Cyclic, TU3

4 Random, TU3
3
Cyclic, TU50
2
1 Random, TU50

0
0 1 2 3 4 5 6 7 8
Number of hopping frequencies

© NOKIA FH_SHOW.PPT
The gain of Frequency Hopping?
The gain of frequency hopping is measured and
simulated to be 5-6dB.

It allows tighter frequency reuse in the network


It allows to install extra TRXs in to cell
By going 4/12 reuse to 3/9 reuse 30% capacity is
gained with the same quality

INTERFERENCE
TIGHTER MORE HIGHER
& FREQUENCY
REUSE TRX/CELL CAPACITY
DIVERSITY

© NOKIA FH_SHOW.PPT
FH support in BSS releases

Feature BSS6 (S6,DF3,B10) BSS7 (S7,DF4,B11)

BB/RF/cyclic/random yes yes

Freeform RF-hopping yes yes


• Frequency sharing on site
• 63 frequencies/MA
• RF-FH up to 4TRX/cell

MAIO Step no yes

Intelligent Frequency Hopping no yes


(IFH = IUO + FH)

BSS8 Feature Candidates (for DE34): Main benefit:


• RTC solution upto 12TRX/cell -> large configurations with FH and minimum antennas
• Wideband solution upto 12TRX/cell -> AFE expansion path for large configurations

© NOKIA FH_SHOW.PPT
FH Implementation

MS does not see


TRX-1 F1(+ BCCH) any difference
BB-FH
F2
Frequency
TRX-3 F3
Dig. RF F1
PSTN F2
BSC F3
MSC
TCSM
Time

TRX-1 F1, F2, F3

RF-FH
TRX-2 BCCH
Dig. RF

BB-FH is feasible with large configurations


RF-FH is viable with smaller configurations

© NOKIA FH_SHOW.PPT
FH sequence generation with MAIO effect
GSM Hopping algorithm
TCH (MAIO=2)
TCH (MAIO=1)
TCH (MAIO=0)
FN/MA/MAIO/HSN

BTS

MA f1 f2 f3 f4 fN-1 fN

MA INDEX 0 1 2 3 N-2 N-1

MAI (0..N-1) + MAIO


MAIO's enable frames (and TCHs) within same hopping group to use different frequencies.

© NOKIA FH_SHOW.PPT
BB hopping

Different MAIOs
for same RTSLs
BCCH timeslot, does not hop.
of TRXs within
the same
RTSL-0 RTSL-1 RTSL-2 RTSL-3 RTSL-4 RTSL-5 RTSL-6 RTSL-7
hopping group

TRX-1 BCCH 0 0 0 0 0 0 0 f1 -> no collisions.

TRX-2 0 1 1 1 1 1 1 1 f2

TRX-3 1 2 2 2 2 2 2 2 f3

TRX-4 2 3 3 3 3 3 3 3 f4

Timeslot 0 of TRXs 2-4 hop over MA(f2,f3,f4). All timeslots 1-7 hop over MA(f1,f2,f3,f4).

© NOKIA FH_SHOW.PPT
Hopping TCHs in BTS
BB-FH case

TDMA frame: 0 1 2 3 4

TRX-1 / f1

TRX-2 / f2

TRX-3 / f3

MAI :
TCH-1 (MAIO 0) 0 2 0 1 2
TCH-2 (MAIO 1) 1 0 1 2 0
TCH-3 (MAIO 2) 2 1 2 0 1

MA list f1 f2 f3 MAI = (S + MAIO) modulo N


S = calculated on frame basis by MS and BTS
using GSM hopping algorithm and
MAI 0 1 2 MA/HSN/FN parameters
N = length of MA list

© NOKIA FH_SHOW.PPT
RF hopping

BCCH TRX, does not hop.

RTSL-0 RTSL-1 RTSL-2 RTSL-3 RTSL-4 RTSL-5 RTSL-6 RTSL-7

TRX-1 BCCH 0 0 0 0 0 0 0 MA = {f1}

TRX-2 0 0 0 0 0 0 0 0

TRX-3 1 1 1 1 1 1 1 1 MA = {f2, f3, f4,..}

TRX-4 2 2 2 2 2 2 2 2

MAIOs have to
be different for
Only one different TRXs
hopping group within the same
hopping group
-> no collisions.

© NOKIA FH_SHOW.PPT
Hopping TCHs in BTS
RF-FH case

TDMA frame: 0 1 2 3 4

TRX-1

TRX-2

MAI / freq. :
TCH-1 (MAIO 0) 0 / f1 2 / f3 0 / f1 1 / f2 2 / f3
TCH-2 (MAIO 1) 1 / f2 0 / f1 1 / f2 2 / f3 0 / f1

MA list f1 f2 f3 MAI = (S + MAIO) modulo N


S = calculated on frame basis by MS and BTS
using GSM hopping algorithm and
MAI 0 1 2 MA/HSN/FN parameters
N = length of MA list

© NOKIA FH_SHOW.PPT
Hard/soft blocking

Hard blocking

The whole radio resource is in use - no more calls can be established


due to lack of free radio timeslots.

Dominates with large reuse factors

Soft blocking

The capacity of individual cells is limited by the level of the interference


rather than the number of TRXs available.

Is dominating with tight reuse patterns.

© NOKIA FH_SHOW.PPT
Fractional Load (FL), RF FH
•“HW load” is 75%
•Fractional load FL is
75 % 25 % 3 TRX / 5 F = 0.6 = 60%
•“Frequency load” is
HW load * FL = 45%

TRX-1 BCCH 0 0 0 0 0 0 0 f1

TRX-2 0 0 0 0 0 0 0 0 f2, f3, f4, f5, f6

TRX-3 1 1 1 1 1 1 1 1 f2, f3, f4, f5, f6

TRX-4 2 2 2 2 2 2 2 2 f2, f3, f4, f5, f6

Active slots Empty slots

© NOKIA FH_SHOW.PPT
FH simulations

REUSE 1/1 1/3 3/9

Blocking type Soft Soft Hard


Max loading 8% 33% 100%
(freq. load)

'Safe' reuse, good


quality can be
With this load the quality may achieved
drop much compared to 3/9
Should work even
PC and DTX should be used without PC and DTX

Effective reuse of 9 is feasible with FH, if the reuse is much


tighter, the quality may be endangered.

© NOKIA FH_SHOW.PPT
Loading of 1/3 RF-FH
Simulation

8
7
6
Dropped call rate [%]

5
4
3
2 2% CDR
1
0
9,94 19,89 33,67 38,99
Load [%]

33% Load
Small increase in loading at soft blocking limit will degrade the network quality badly.

© NOKIA FH_SHOW.PPT
What is a feasible reuse?
Load control needed
Erl/cell Safe to implement
in practice

+91%

+43% +44% +39%

reference

CALL DROP RATE: 0% 0.2% 2% 2% 2%

REUSE FACTOR: 12 9 6 3 1
(Freq. Allocation reuse) (no hopping)

BLOCKING TYPE: Hard Hard Soft Soft Soft

© NOKIA FH_SHOW.PPT
The effect of Power Control and DTX
downlink simulation

Reuse 3/9, TU 3km/h Reuse 3/9, TU 50km/h

GAIN: GAIN:
PC on, DTX off 1.4 dB PC on, DTX off 1.0 dB
PC off, DTX on 2.3 dB PC off, DTX on 2.3 dB
PC on, DTX on 3.7 dB PC on, DTX on 3.5 dB

C/I improvement

The gain achieved with PC is lower than the DTX gain.

© NOKIA FH_SHOW.PPT
RXqual vs. FER
Measurement
30
40

25 35

keskimääräinen FE R[%]
30
20
Average FER

25

15 20

15
10
No hop Hopping
10
5
5

0 0
0 1 2 3 4 5 6 7
0 1 2keskimääräinen
3 RX QUA4L 5 6 7
Average RXqual

With hopping adequate speech quality can be achieved with low RxQual.

© NOKIA FH_SHOW.PPT
Nokia FH Solutions

1. Solutions for large bandwidth and maximised capacity


• BB-FH with 3/9 reuse pattern
• BB-FH with mixed reuse

2. Solutions for narrow bandwidth


• RF-FH with 1/3 (or 3/9 ) reuse using frequency sharing and
fractional loading

3. Solutions for easy frequency planning


• RF-FH with 1/1 reuse pattern using frequency sharing and
fractional loading

© NOKIA FH_SHOW.PPT
BB-FH with 3/9
Number of hopping frequencies needed
= 9 x number of hopping TRXs/cell Baseband
hopping
BCCH, reuse 15
0 1 2 3 4 5 6 7
BTS BTS BTS BTS
TRX-1, reuse 9
0 1 2 3 4 5 6 7
3/9 CLUSTER:
• 3 sites
• 9 cells BTS BTS BTS TRX-2, reuse 9
0 1 2 3 4 5 6 7

TRX-3, reuse 9
BTS BTS BTS BTS 0 1 2 3 4 5 6 7

HOPPING FREQUENCY AN EXAMPLE OF HOPPING WITHIN ONE


GROUP CELL WITH 4TRXs

Method with proven performance, all RTSLs have a reuse of 9 or more.

© NOKIA FH_SHOW.PPT
RX Quality Distribution in one
3/9 FH Network
TCH call drop ratio: 1.71 %
TCH access probability: 99.98 %

100
88,41
80
60
%
40
20 1,79 1,96 2 1,85 1,52 1,6 0,87

0
q0 q1 q2 q3 q4 q5 q6 q7
UL RX quality class

© NOKIA FH_SHOW.PPT
BB-FH with Mixed Reuses

Baseband
hopping
BCCH, reuse 15
0 1 2 3 4 5 6 7

TRX-1, reuse 9
0 1 2 3 4 5 6 7

TRX-2, reuse 9
0 1 2 3 4 5 6 7

TRX-3, reuse 6
0 1 2 3 4 5 6 7

Planning method where the individual TRXs within a cell RTSL-0 has tighter reuse than others,
are allocated frequencies of successively tighter reuse which increases call drop rate

© NOKIA FH_SHOW.PPT
BB-FH Capacity Gain

35 %
30 %
The gain is
25 %
approximately 30%
20 %
BB-FH is not suitable
with narrow BW 15 %
10 %
Reuse 3/9
5%
0%
30 33 36 39 42 45 48 51 54 57 60
Channels

© NOKIA FH_SHOW.PPT
RF-FH with 3/3 Reuse and Frequency Sharing

MA1 = f1, f4, f7 F6 F5


MA2 = f2, f5, f8
MA3 = f3, f6, f9 F9 F8
F3 F2
Initially 3/9 Cluster, F1 F3
3/3 with Frequency sharing
F9 F8
F4 F6
F3 F2
F7 F9
F6 F5
F6 F2 F7 F3
F9 F5 F1 F6
F3 F8 F4 F9 Time
F1 F6
F4 F9
F7 F3

As the hopping is random and


sites not synchronised, adjacent
channels are used at times
in neighbouring sites

© NOKIA FH_SHOW.PPT
RF-FH with 'reuse 1/3'
using fractional loading
Band allocation:

BCCH

MA_1 MA_1
MA_2 MA_2 MA lists are yield e.g. by:
MA_3 MA_3 dividing the hopping frequencies into 3
MA_1 MA_1 groups and then adjusting the lengths of
MA_2 MA lists based on interference and
MA_3 number of TRXs in individual cells
MA_1

Effective reuse > 3 as fractional loading used

© NOKIA FH_SHOW.PPT
RF-FH with reuse '1/1'
using fractional loading and MA list sharing

MAIO Offset Hopping frequencies


MAIO Offset + Step
Band allocation: Band allocation:

BCCH BCCH Hopping Freq's

MA list MA list

MA list and BCCH need planning No need for MA list planning

MA list possibly shorter -> reduced gain BCCH frequencies planned as usual

© NOKIA FH_SHOW.PPT
Application Areas of FH Schemes
BB BB RF RF 1) RF
3/9 Mixed '3/3' '1/3' '1/1'
Maximised capacity

Narrow

BANDWIDTH
Medium

Large

Easy Planning

Easy Implementation 2)

1) Heuristical planning assumed


2) No load control needed

© NOKIA FH_SHOW.PPT
Possible Evolution Paths

Small BW Medium BW Large BW Existing IUO


6 MHz 8 MHz 10 MHz network

Phase

S6 RF BB hopping BB hopping
FH (AFE used) IUO
hopping

S7 RF RF BB hopping
IFH hopping hopping

© NOKIA FH_SHOW.PPT
Frequency Hopping Evolution Path
2. gen. BTS 3. gen. BTS

RTC AFE
RTC (or WBC)

•min 3 TRX •min 2 TRX


•DE34: max BB FH RF FH
•max 12 TRX in
6 TRX in BSS7 BSS7

© NOKIA FH_SHOW.PPT
BTS Distribution in FH Network
one example
Network having both
2nd gen and 3rd gen BTS’s

2nd gen BTS’s

Macro layer BTS’s with large


BB FH configurations (min. 3 TRX’s).
BB hopping used.

3rd gen BTS’s


Macro layer BTS’s with small
RF FH configurations and Micro
layer BTS’s (min. 2 TRX’s).
RF hopping used.

© NOKIA FH_SHOW.PPT
Nokia RF FH Trial

© NOKIA FH_SHOW.PPT
Gain of Frequency Hopping?

The gain of frequency hopping is


measured and simulated to be 6-
8dB

• It allows a tighter frequency reuse in


the network
• It allows to install extra TRXs in a
cell
• By going from 4/12 reuse to 3/9
reuse, 30% capacity is gained with
the same quality

© NOKIA FH_SHOW.PPT
Listening Tests of RXQUAL and FER
Subjective quality, laboratory tests

Steady quality/FER value (fast mobile or frequency hopping)

RXqual FER
0-4 good 0 - 4% good
5 slightly degraded 4 - 15% slightly degraded
6 degraded 15 - 35% degraded
7 useless >35% useless

•RXQUAL 6 shows rapid transition: low 6 is almost OK but high 6 is bad


•FER has finer resolution

© NOKIA FH_SHOW.PPT
HO Parameters

Inter cell HO happened mainly based on power budget or


quality.

HO period for power budget was 2 SACCH blocks equals


to 1 s.

Handover Type Margins Thresh. Averages


UL DL UL DL
Power budget HO 4 dB
Level HO 3 dB
Quality HO 0 dB 6 7 4/6 1/4

© NOKIA FH_SHOW.PPT
Power Control Parameters

Power control was based mainly on quality

Power increasing and decreasing step size was 2 dB with 1 s interval

Power control attenuation range was from 0 to 20 dB, but it was


increased from 0 to 30 dB in the one MA-list cases

MS power level was optimized after HO so that UL_RX_lev was -79


dBm
Beginning 1/1 reuse
UL DL UL DL
QL threshold
Power Increase 3 4 2 4
QL threshold
Power Decrease 1 1 2 2
Amount of Averages 2/3 1/1 2/3 1/1

© NOKIA FH_SHOW.PPT
Frequency Allocation Strategy, 1/3
Reuse
• In RF Hopping the total load was adjusted for every cell
by using the fractional loading.
• The number of needed frequencies and TRX's was
calculated separately for each cell based on the busy
hour traffic.
• At least three frequencies were allocated to each cell to
ensure an adequate hopping gain.
• Two different scenarios for frequency allocation were
used:
• 3 fixed MA lists
• heuristically planned (C/I matrix) different lists

© NOKIA FH_SHOW.PPT
Frequency Allocation Strategy, 1/1
Reuse

• Two different scenarios for frequency allocation were


used:
• The number of needed frequencies was calculated separately
for each cell based on the busy hour traffic. At least 13
frequencies were allocated to each cell to ensure an adequate
hopping gain.
• 15 frequencies were assigned to each cell

© NOKIA FH_SHOW.PPT
General Observations from Frequency Hopping
Trial

• Call drop rate due to interference was nearly zero


• UL&DL Quality in OMC statistics got worse when UL/DL DTX was
on
• NMS/X statistics showed improved quality when UL/DL DTX was
on
• DL power control improved DL quality
• Inter-sector orthogonality improved DL quality
• Misbehaviour of some mobile phones with hopping was noticed
• In FH case about 2 x more bad quality still resulted in the same
subjective speech quality

© NOKIA FH_SHOW.PPT
Quality Analysis

• power control and RF FH with smaller reuses


decreased the amount of samples in quality class 0
and increased the number of samples in quality
classes 3-5; the effect was much stronger in DL
• OMC quality statistics shows worse quality when DTX
is activated, because subRXQual samples used in the
statistics
• NMS/X quality statistics shows improved quality when
DTX is activated
• DL DTX caused some quality problems for some
mobiles
• UL quality got better than DL quality with 1 reuse

© NOKIA FH_SHOW.PPT
Conclusions of FH Trial
•an effective reuse drop from 14-> 8 became feasible
with 3 frequency allocation reuse and heuristic
planning, used together with UL/DL power control and
UL DTX -> 40% capacity gain with worse quality
•the heuristic planning of hopping frequencies based
on the actual interference probability relations
provided the best capacity in the trial

•the usage of only one MA-list was found to be feasible


and it clearly outperformed the pure 1/3 case where
the band was divided into 3 fixed MA lists
•difficult to estimate the subjective speech quality
when DTX used because of subRXQUAL values used
as a measure
•FER-Quality mapping helps to estimate the subjective
speech quality with different FH schemes
© NOKIA FH_SHOW.PPT
Conclusions of FH Trial
•simulated frequency loads about matched to the
trialled ones

•new requirements for KPIs needed when FH used in


the network; FER -> Quality mapping used as a basis

•DL DTX still not recommended to be used because of


mobile and network? problems

© NOKIA FH_SHOW.PPT
'Freeform RF Hopping'
(or 'MAIO Management')

© NOKIA FH_SHOW.PPT
Freeform RF Hopping
• Introduced in BSS 6 and BSS7

• Allows for flexible RF hopping:


• enables Frequency Sharing i.e. sharing an MA list between the sectors at
same site
• longer MA lists possible
• minimised interference

• In BSS6 MAIO Offset was introduced:


• separate starting MAIOs for co-sited cells

• MAIO Step parameter in BSS7


• when used together with MAIO offset, no successive MAIOs will ne allocated for
TCHs sharing the same MA list.

© NOKIA FH_SHOW.PPT
FH sequence generation
GSM Hopping algorithm

Downlink:
FN/MA/MAIO/HSN

BTS

MA f1 f2 f3 f4 fN-1 fN

INDEX 0 1 2 3 N-2 N-1

MAI (0..N-1)

© NOKIA FH_SHOW.PPT
FH sequence generation with MAIO effect
GSM Hopping algorithm
TCH (MAIO=2)
TCH (MAIO=1)
TCH (MAIO=0)
FN/MA/MAIO/HSN

BTS

MA f1 f2 f3 f4 fN-1 fN

MA INDEX 0 1 2 3 N-2 N-1

MAI (0..N-1) + MAIO


MAIO's enable frames (and TCHs) within same hopping group to use different frequencies.

© NOKIA FH_SHOW.PPT
BSS6 Single MA/HSN per Site - Using MAIO Offset
MA list can't include
HSN same for all sectors MA = f1, f3, f5, f7,....
adjacent frequencies

MAIO, same for all


Sector HSN MO TRX
RTSLs within the TRX
1 TRX-1 BCCH, not hopping
N 0 TRX-2 0
TRX-3 1
TRX-4 2 Only half of the band
2 TRX-5 BCCH, not hopping allocation can be included
N 3 TRX-6 3 into one MA list
TRX-7 4
TRX-8 5
3 TRX-9 BCCH, not hopping
N 6 TRX-10 6
TRX-11 7
TRX-12 8

Operator can set the lowest


MAIOs for the cells

© NOKIA FH_SHOW.PPT
BSS7

Benefits of MAIO Offset

MAIO Offset has the following benefits:

frequency sharing between the cells inside the site


- > longer MA lists possible

hopping possible also with narrow BW and


short MA list

© NOKIA FH_SHOW.PPT
BSS7 Single MA/HSN per Site - with MAIO Step

MA = f1, f2, f3, f4,.... MA list can include


HSN same for all sectors adjacent frequencies

MAIO, same for all


Sector HSN MAIO Offset MAIO step TRX
RTSLs within the TRX
Band allocation:
1 TRX-1 BCCH, not hopping
1 N 0 2 TRX-2 0 BCCH Hopping Freq's
TRX-3 2
TRX-4 4
2 TRX-5 BCCH, not hopping
2 N 6 2 TRX-6 6
Nor co-channels
TRX-7 8
neither adj. channels
TRX-8 10
used simultaneouly
3 TRX-9 BCCH, not hopping
if number of
3 N 12 2 TRX-10 12
frequencies >
TRX-11 14
2*number of TRXs
TRX-12 16

Operator can set the lowest Operator can also set the
MAIOs for the cells MAIO step size

© NOKIA FH_SHOW.PPT
BSS7

MAIO Step

MAIO Step will bring the following benefits:

one MA list per site

one MA list can contain a continuous band

no risk of co-channel nor adjacent channel being used


simultaneously within a site

single MA/HSN possible -> only BCCH frequency planning

MAIO step has to be used together with MAIO offset

© NOKIA FH_SHOW.PPT
RF-FH with reuse '1/1'
using fractional loading

MAIO Offset Hopping frequencies


MAIO Offset + Step
Band allocation: Band allocation:

BCCH BCCH Hopping Freq's

MA list MA list

MA list and BCCH need planning No need for MA list planning

MA list possibly shorter -> reduced gain BCCH frequencies planned as usual

© NOKIA FH_SHOW.PPT
INTELLIGENT UNDERLAY OVERLAY
( IUO )

© NOKIA FH_SHOW.PPT
What is ‘capacity’ in a cellular
system ?
• Objective : to support as many subscribers as possible
• Limitation of a ‘radio’ system : number of radio frequencies is
‘finite’
• GSM
•124 pairs of frequencies
•every frequency ‘pair’ can support a max. of 8
simultaneous calls

ONLY 124*8 = 992 SIMULTANEOUS CALLS CAN


BE SUPPORTED

How to overcome this limitation :


The answer is ‘reuse’, i.e. to use the same set of frequencies
over and over again.

© NOKIA FH_SHOW.PPT
What is ‘Reuse’?
Cell A •Entire network area divided into ‘cells’
•Each cell uses a given set of frequencies e.g.
Cell A uses frequencies f1 and f4
•Ideally, it should be possible to use both f1
and f4 in all the cells, viz, A, B, C, ...J,K,L.
f1& f4 This means a ‘reuse’ of 1, i.e every cell uses
the same set of frequencies
Cell B •Practically, it is not possible since these
frequencies ‘interfere’ at such short
Cell C distances
Cell D Cell G •So, in real life it is possible to use the same
set ,i.e f1 and f4 in Cell M. Since Cell M is 12
Cell E Cell H cells apart from Cell A, the reuse here is 12.
•The set of Cells A....L is called a ‘cluster’.
Cell M Cell F Cell I
e.g. blue cells form a cluster
f1& f4 Cell J
•All the cells within a cluster have to use
Cell K ‘unique’ frequencies to avoid interference
•All the frequencies used in blue cluster can
Cell L be re-used in yellow cluster

Base Station Site

© NOKIA FH_SHOW.PPT
Capacity in GSM depends on..?
• Spectrum
• more the number of frequencies, more the capacity
(max. spectrum possible for GSM = 25MHz or 124 freq. pairs)
• Reuse
• smaller the reuse, more the number of carriers per cell
(Let us take spectrum allocated S = 6.2 MHz. or 31 freq.
Case I : Reuse =12
No. of freqs. possible per cell = 31/12 = 2.58
Case II : Reuse = 9
No. of freqs. possible per cell = 31/9 = 3.44
Case III: Reuse = 1
No. of freqs. possible per cell = 31/1 = 31)
• Area per cell
• smaller the area per cell, more the capacity
(Given network area = 96 sq. km., reuse = 12
Case I, Average cell area = 2 sq.km., Cluster area= 12*2 =24sq.km.
No. of clusters = 96/24 = 4 (i.e. same set of freqs. can be reused in the network
area 4 times)
Case II, Average cell area = 1 sq.km., Cluster area = 12*1 =12sq.km.
No. of clusters = 96/12 = 8 (i.e. same set of freqs. can be resued in the network
area 8 times => more capacity))

© NOKIA FH_SHOW.PPT
How to achieve capacity increase?
S= total spectrum available
Capacity in GSM proportional to: S n=re-use factor
n*A A=area/cell

Ways to increase capacity :


(a) Increasing Spectrum :
More GSM bandwidth or GSM 1800 band : discretion of the regulator

(b) Decreasing Cell Area :


Microcellular solution : large number of sites; very expensive

(c) Reducing reuse Factor


Intelligent Underlay Overlay , IUO (Software capacity)

© NOKIA FH_SHOW.PPT
What exactly is IUO?

Regular layer (reuse = 12) Regular layer (reuse = 12)


Regular layer (reuse = 12) Regular layer (reuse = 12)

• Two layer network - regular and super


• Regular layer with regular reuse (e.g. 12); super layer with aggressive reuse (e.g. 7)
• Regular layer takes care of coverage
• Super layer provides capacity
• ‘Dominance’ region of super layer controlled by ‘interference’
• Calls close to the base station are handled in super layer
• Call handed over from super layer to regular layer when ‘interference’ becomes excessive

<-------- Interference region ---------->

Super layer (reuse = 7) Super layer (reuse = 7)

© NOKIA FH_SHOW.PPT
What is so ‘intelligent’ about IUO?
Case I : Indoors

• Conventional Underlay Overlay • Intelligent Underlay Overlay

Measured signal level, Co (outdoor) = -75 dBm Measured server, Co = -75dBm


Building penetration loss, B = 15 dB Measured interferer, Io = -90dBm
Defined Condition for handover from ‘capacity’ layer Difference between Server and interferer
to ‘coverage’ layer is, Signal level worse than -80 = Co/ Io = Co - Io =15dB
dBm Defined conditions for handover :
Measured Signal level, Ci (indoor) = Co-B = -90 C/I better than 14dB => Super layer
dBm
C/I worse than 11dB => Regular layer
Since, Ci is worse than -80 dBm
Indoors :
=> Handover to coverage layer
Measured server, Ci = Co - B = -90dBm
Measured interferer, Ii = Io - B = -105dBm
Ci/ Ii = Ci - Ii =15dB
=> Call stays in super layer

The intelligence of IUO is in the dynamic measurement of interference of every Mobile.


=> Indoors, the achieved capacity increase is more than that with
conventional underlay overlay solutions

© NOKIA FH_SHOW.PPT
‘Intelligence’ of IUO (Contd.)
Case II : Interfering signal has line of sight
Interferer Carrier

• Conventional Underlay Overlay • Intelligent Underlay Overlay

Measured signal level initially,C1 = -81 dBm Measured carrier initially, C1 = -81 dBm
Measured signal level after a while,C2 = -75 dBm Measured interferer initially, I1 = -97dBm
Defined Conditions for handover: Difference between carrier and interferer
Carrier better than -78dBm =>capacity layer = C1/ I1 = C1 - I1 =16dB
Carrier worse than -80 dBm =>coverage layer Defined conditions for handover :
Since interference estimation not possible before the C/I better than 14dB => Super layer
handover, call is transferred to capacity layer C/I worse than 11dB => Regular layer
=> Call is most probably lost due to ‘strong’ => Call in super layer
interferer which has line of sight.
After a while,
Measured carrier, C2 = -75dBm
Measured interferer, I2 = -85dBm
C2/ I2 = C2 - I2 =10dB
=> Call handed back to regular layer which has
cleaner frequencies

=>IUO is better suited to handle typical ‘dense urban’ environments

© NOKIA FH_SHOW.PPT
What is the capacity increase achieved?

• Example case , spectrum = 6.2 MHz., 31 freqs

Config. Conventional IUO


Reuse 12 Regular layer = 12
Super layer = 7
No. of Regular TRXs/cell 31/12 = 2.58 24/12 = 2
No. of Super TRXs/cell 0 7/7 = 1
Total no. of TRXs/cell 2.58 3
Traffic/cell 11.5 Erl 14.9 Erl

Achieved capacity increase = 30%

© NOKIA FH_SHOW.PPT
Intelligent Underlay Overlay
increases Operational Efficiency
• Easy frequency planning of capacity expansion when using IUO
• No need for new sites -> rapid implementation
• Easy quality and capacity adjustment
• IUO does not prevent using other capacity enhancement features
• Security of maintaining the quality of service as dynamic C/I measurements
are done continuously with IUO
• Various IUO planning and implementation services available incl. turnkey
• Available today - successfully implemented in around 20 networks

=>IUO results in :
•Reduced Cost of Ownership
•Increased Operational Efficiency

© NOKIA FH_SHOW.PPT