5G Ultra-Dense

Femtocell
Network Access:
Issues and
Challenges

Prof. Dr. Mahamod

Ismail
mahamod@ukm.edu.m
y

2016 Dr.MBI@UKM

Outline
Introduction
Ultra Dense (Femtocells)
Network
Interference Management
Related Research
Conclusion

2016 Dr.MBI@UKM

Abstract
Due to the popularity of data-hungry devices, like smart phones and tablets, the mobile operators
towards the Fifth Generation (5G) are driven not only to increase their network capacity but also to
boost the life of such battery operated wireless smart devices. One of the most feasible solutions to
cope with this is the evolution from traditional macro-cell deployments to network densification. Small
cells are seen as the best match for network densification, as they can be opportunistically deployed in
the highly irregular way in hot spots. In consequence, the deployment of dense femtocell networks
(DFNs) can be seen as an efficient spectrum utilization using new spectrum bands but subjected to
inter-cell interference coordination particularly to cell-edge macro-users. Moreover, the decentralized
architecture of the femtocell networks and the uncertainty in terms of the number and location of
femtocell base stations increases the complexity of power control technique. This presentation will
discuss inter-femtocell interference scenarios in a typical femtocell cluster together with an overview of
some of the available interference management techniques. An interference mitigation schemes in
indoor LTE-A femtocell network using a new simple fractional path-loss compensation power control
technique utilizing local HeNBs information while maintaining QoS been proposed. Since densification of
femtocells in an uncoordinated manner can lead to severe inter-femtocell interference, especially when
they share the same channel and operate in Closed Subscriber Group (CSG) mode, a new Semi-virtual
Clustering Scheme (SVCS) has been investigated. This scheme exploits users status to categorize
femtocells into the victim, aggressor or neutral and each victim femtocell is then partitioned into two
virtual cells. Moreover, the scheme adapts dynamically to the status of each femtocell in the cluster,
being either a victim or a safe femtocell and smartly estimates the proper partitioning (time or
frequency) of resources within the defined clusters.

2016 Dr.MBI@UKM

Introduction

Source: Qualcomm 2013

2016 Dr.MBI@UKM

3GPP Release-12 Onwards

MTC Machine-Type Communications

eMBMS - Evolved Multimedia Broadcast/Multicast Service
D2D Device-to-Device

2016 Dr.MBI@UKM

Introduction

Source: Roberts 2015

2016 Dr.MBI@UKM

Fifth Generation (5G)

High Capacity

High QoE

Efficiency

Latency < 1 ms

High Quality

User throughput ~ 1 Gbps

Low Latency

Avoid capacity crunch

with vast number of IoT
devices

High Throughput

Introduction

Cost efficient high density

small cell capacity and
energy efficient

Long Battery Life

Energy efficiency (up to

10 years)

Source: Roberts 2015 & Benn 2014

2016 Dr.MBI@UKM

Fifth Generation (5G)

Introduction

5G

is a Special Generation (Tafazoli 2015) Always

Sufficient Rate to give users the perception of infinite
capacity.
One way of adding capacity in 5G is by densifying the
network and basically means adding more cell sites
strategically placed in capacity-strained areas (urban
areas and large public venues), add more capacity where
it is most needed and also help offload traffic from
surrounding sites.
Heterogeneous network (HetNet) deploy compact and
low-powered base stations term as Small Cells, that
include Femto cells, Pico cells, Micro cells and Metro cells.

2016 Dr.MBI@UKM

Introduction
Heterogeneous Network
(HetNet)

2016 Dr.MBI@UKM

Small Cell Networks

Introduction

10

2016 Dr.MBI@UKM

Cell Evolution

UDN

Nokia

(2015) - An ultra dense network (UDN) is one with

sites on every lamp post or with indoor sites placed
within 10m of each other. UDNs will cover most urban
indoor and outdoor areas with small cells providing cell
edge data rates of 100 Mbps to everyone.

DAS - Distributed Antenna Systems

LAA - Licensed Assisted Access
C-RAN Centralized Radio Access Network

LWA - LTE Wi-Fi Aggregation

LP RRH Low Power Remote Radio Host

11

2016 Dr.MBI@UKM

Ultra Dense Networks

(UDN)

Zahid Ghadialy 2015

UDN

12

2016 Dr.MBI@UKM

UDN Scenarios

Hao Peng 2015

UDN

13

2016 Dr.MBI@UKM

Cell Properties

NIST 2016

UDN

14

2016 Dr.MBI@UKM

Coverage vs. Capacity

UDN

15

2016 Dr.MBI@UKM

Small Cell (SC) Issues

UDN

Dense

and ad hoc deployment -> new

network models
How to manage interference?
Key to successful deployment of SC
How can SC co-exist with the main stream
wireless system?
Most critically, mobility and handover
What is the best backbone to support SC?
How can we handle dense deployment of?

16

2016 Dr.MBI@UKM

Challenges

UDN

Propagation

and system modeling: - new propagation

models are needed since most channel models assume
independent fading between links:
Tower-to-ground BS at 30 m and above.
Device-to-device links ground-to-ground (few model exists)
Picocells on lightpoles
Indoor femtocells placed on tabletops or mounted on walls or ceilings
will be < 10 m above ground.
In UDN, this assumption does not hold because there will be significant
correlation between links, necessitating the study of joint links.

Interference

mitigation:

Coordination algorithms for synchronizing cells, managing power, and

selecting carriers needs to be examined in terms of their inaccuracy
(timing offsets, etc.)
Determining the limits of densification (i.e. how much capacity reuse
is possible before interference or channel effects or overhead causes
saturation) is an important issue that needs to be addressed.

17

2016 Dr.MBI@UKM

Challenges
Cell-Edge

UDN

Effect

Mobility and interference are negatively affected and

ideal backhaul becomes more and more difficult.
Solution (ZTE):
1.
2.

Smooth cell virtualization with hybrid control mechanism

Self-Backhaul cost effective, guaranteed link quality and
provide traffic offloading

18

2016 Dr.MBI@UKM

UDFNs

UDFNs
UDFN

- Ultra Dense Femtocell Network

Subscribers are using more network resources
than ever before and the demand for more
resources is tremendously increasing. Thanks to
SMART DEVICES
Network operators have to act to meet this
demand:

Buy more spectrum.

Make spectrum more efficient.
Move to network densification.
19

2016 Dr.MBI@UKM

UDFNs
Network

UDFNs
Densification Means:

High concentration of mobile users.

High concentration of cell sites :
Add more capacity where it is most needed, e.g.
capacity-strained areas (urban, large public venues).
Meet the expectation of end users.

Small

cells are seen as the best match for

network densification.
Femtocell densification has a great attraction
in 3GPP since Release-10.
20

2016 Dr.MBI@UKM

Challenges

UDFNs

As femtocells are deployed by end users in an

uncoordinated manner, many problems arise; one
amongst others is:

Downlink interference can become trickier.

More disperse definition of aggressor and victim
femtocells .

Identifying victim and aggressor femtocells is

prerequisite to achieve an effective interference
management scheme.
Need
to exploits users status to categorize
femtocells into either victim, victim-aggressor ,
aggressor or neutral.

21

2016 Dr.MBI@UKM

Interference

22

2016 Dr.MBI@UKM

Interference
SIC Successive Interference
Cancellation
PIC Parallel Interference Cancellation
MSIC Multistage SIC
MUD Multi-User Detection

Source: Zahir et al. 2013

Interference Mitigation; Resource Allocation, Access Control, Load Balancing, Cognitive

23
2016 Dr.MBI@UKM
Approach, Q-Learning (Ahmed et al. 2014)

Interference Scenarios

Interference

24

2016 Dr.MBI@UKM

Interference Scenarios

CSG - Closed Subscriber Group

HeNB Home eNode-B

Interference

25

2016 Dr.MBI@UKM

Inter-cell Coordination

Interference

26

2016 Dr.MBI@UKM

Inter-cell Coordination

Interference

27

2016 Dr.MBI@UKM

Inter-cell Coordination

Interference

28

2016 Dr.MBI@UKM

Related Research

Partially Compensated

Power Control Technique for

LTE-A Macro-Femto
Networks

The minimum transmit power level

is calculated based on partial
compensation for the pathloss to
mitigate interference to
neighbours.

29

2016 Dr.MBI@UKM

Related Research

The proposed scheme reduces the outage probability of macro-UEs,

while maintaining good SE for HUEs.

30

2016 Dr.MBI@UKM

Related Research
Semi-Virtual Clustering for Mitigating Inter-Cell Interference in
UDFNs

Densification of femtocells in an uncoordinated manner can lead to severe interfemtocell interference when sharing the same channel and operate in Closed
Subscriber Group (CSG) mode.

Identifying victim and aggressor femtocells is prerequisite to achieve an effective

interference management scheme.

A new approach using semi-virtual clustering scheme (SVCS) of femtocells in an

ultra-dense environment and exploiting users status to categorize femtocells into
the victim, aggressor or neutral.

Each victim femtocell is then partitioned into two virtual cells serving: (i) victim
users and (ii) safe users.

Two new virtual clusters are then formed. One cluster is for the first virtual cells,
but the latter one retains its original physical cluster, while the other cluster is for
their aggressors. Accordingly, aggressor femtocells become neutral, thereby
ensuring minimum inter-femtocell interference

31

2016 Dr.MBI@UKM

Aggressors for a UE can be defined as the plurality of interferers sorted from the strongest to the weakest. The dominant interferer (DI) is the strongest amongst all interferers. The
as:

UDFN Interference

Related Research

Based

on users status , in
any cluster, a femtocell can
be classified as any type:
Victim femtocell (F4, F7, F10).
Victim-aggressor femtocell (F1).
Aggressor femtocell (F2, F5, F8, F9).
Neutral femtocell (F3, F6).

1.
2.
3.
4.

Classification

of a UE to be
critical is based on two
metrics:
1.
2.

Dominant interference ratio (DIR) <3dB.

Signal to Dominant interference (DI)
ratio- <10 dB.

DIR

I strongest

i strongest

32

Ii N

2016 Dr.MBI@UKM

Related Research
Concept of
SVCS
Type 2 and Type 3 is the main focus for SVCS to be an efficient
interference management scheme.
On the other hand, the critical UEs in Type 1 and Type 2 are the
targets for the performance evaluation of this interference
mitigating technique.
The Type of any femtocell in a cluster is time variant.
More specifically, the UE location in femtocell and the distance to
its neighbors define the femtocells Class.
SVCS dynamically adopts to this variations.
SVCS smartly estimates the proper partitioning of radio resources
(time or frequency) within the defined clusters according to the
Class of femtocells..
33

2016 Dr.MBI@UKM

Scenario

Related Research

For

our simulation, the focus is on 3GPP

Scenario 3, where it models a dual block of
two or more floors; comprising 10 small
residential apartments each of 10x10m
area.
This scenario imitates a typical urban
hotspot with high density of uncoordinated
deployed CSG femtocells and the risk of
inter-cell interference is at its extreme.
We take
into consideration a more
realistic UDFN deployment scenario, with a
quad-strip model with two floors.

34

2016 Dr.MBI@UKM

Related Research
Simulations assumptions
Parameter
Typical value
Carrier bandwidth
10 MHz
Carrier frequency
3.5 GHz
Transmit power
24 dBm
Femtocell antenna height
6m
UE antenna height
1.5m
Antenna gain and losses
5dBi
UE speed
3km/h
Femtos per floor
2-10
Users per femtocell
1-10
Mode of operation
CSG
Inner wall penetration loss
5dB
Outer wall penetration loss
23dB
Floor penetration loss
0dB, 18.3dB
Path loss model
ITU InH
Shadowing standard deviation
LOS 3dB,
NLOS 4dB
Number of available RBs
18
35

2016 Dr.MBI@UKM

Results

Related Research

Figure shows the number of critical and non- critical UEs compared to
the total number of UEs as a function of the number of UEs per
femtocell.
The average percentage is 25 and 75 for victim and safe UEs
respectively.

36

2016 Dr.MBI@UKM

Related Research

Figure shows the percentage of the different Classes of

femtocells as a function of number of femtos per floor.
Significant variations of Classes occurs in low density of
deployment, <5 femtos per floor and as deployment gets denser
(>5 femtos per floor), the variation gets lesser because femtos
get closer to each other and the majority of victim femtocells act
as well as aggressors to their neighbors.
v-ag: victim-aggressor
v: victim
ag: aggressor
neut: neutral.

37

2016 Dr.MBI@UKM

Throughput & SINR

Fig. 3. SINR performance of SVCS and WCS models

WCS Without Clustering Scheme

SCCS Semi-Virtual Clustering Scheme

Related Research

Fig. 4. Mean throughput per user as a function of the

number of users per femtocell

38

2016 Dr.MBI@UKM

Throughput & SINR

Fig. 5. Mean throughput per user as a function of the

number of femtocells per floor

Related Research

Fig. 6. Mean throughput per femtocell as a function of the

number of femtocells per floor

Simulation results have substantiated the significant performance

improvements provided by SVCS over distributed non-clustered systems,
in terms of interference mitigation, higher received SINR and enhanced
users mean throughput.
39

2016 Dr.MBI@UKM

Related Research

Efficient Adaptive Handover

Techniques Over Coordinated

Contiguous Carrier Aggregation
Deployment Scenario In LTEAdvanced System

CC-CADS deployment scheme is using

two contiguous CCs with different
beam orientation for each carrier to
enhance the coverage of the eNB

40

2016 Dr.MBI@UKM

Related Research
The average RSRP, SINR, spectral efficiency and outage probability in CCCADS scenario are significantly better compared to the typical CADSs.
Empirical CDF

8
7

CADS-1
CADS-2
CADS-3
CC-CADS

0.9
0.8

Average SINR [dB]

0.7

0.6
0.5
0.4

CADS-1
CADS-2
CADS-3
CC-CADS

-1

0.1

-2
-54.5

-56

-55

-54
-53
-52
-51
Average Serving RSRP [P r (dBm)]

-50

-49

-54

0.35

-53.5
-53
-52.5
-52
Average Serving RSRP [dBm]

0.3
CADS-1
CADS-2
CADS-3
CC-CADS

0.9
0.8
0.7
0.6
0.5
0.4
0.3

-51.5

-51

40km
60km
80km
100km
120km
140km

Empirical CDF

0.25
0.2
0.15
0.1
0.05

0.2

0.1
0
2.2

0.2

0
-57

0.3

Average Outage Probability [ < thr ]

CDF Probability of Users RSRP [P

> Q rxlevmin]

CDF of Spectral Efficiency Probability

2.4

2.6
2.8
3
3.2
3.4
3.6
3.8
Average UEs Spectral Efficiency [bps/Hz]

4.2

CADS-1

CADS-2
CADS-3
CC-CADS
Carrier Aggregation Deployment Scenarios

41

2016 Dr.MBI@UKM

Related Research
Interference Mitigation Strategies for Co-Existence
Among 5G Heterogeneous Networks

42

2016 Dr.MBI@UKM

Conclusion
Among

others, interference mitigation and capacity

enhancement are two important issues to be resolved
before 5G deployment.
Cell densification can support high traffic load with
proper interference mitigation scheme such as power
control, coordinated continuous carrier aggregation
and semi-virtual clustering scheme.
There is significant performance improvements
provided by SVCS over distributed non-clustered
systems, in terms of interference mitigation, higher
received SINR and enhanced users mean throughput.

43

2016 Dr.MBI@UKM

References

Miranda, J.P. 2014. Interference Mitigation & Massive MIMO for 5G: Summary of
CPqDs Results.
Sawsan Ali Saad, Mahamod Ismail, and Rosdiadee Nordin. 2015. Partially
Compensated Power Control Technique for LTE-A Macro-Femto Networks. Lecture
Notes in Electrical Engineering.315:331-341.
Shayea, I., M. Ismail, R. Nordin & H. Mohamad 2014. Handover Performance over a
Coordinated Contiguous Carrier Aggregation Deployment Scenario in the LTEAdvanced System. International Journal of Vehicular Technology 2014(15):1-15.
Tafazolli, R. 2015. 5G: Special Generation. 5G IMT Seminar, Malaysia
Konstantinos Dimou. 2013. Interference Management Within 3GPP LTE-Advanced.
Hao Peng. Ultra Dense Network in 5G. ZTE Technologies. 2015
Phil Roberts, 5G is this the technology that will deliver the ultimate mobile
experience? 2015 (http://telecom.com)
Nokia. Ultra Dense Networks (UDN) White Paper. 2015
NIST. Ultra Dense Networks. 2016

44

2016 Dr.MBI@UKM

References

Qian Li,Huaning Niu, Apostolos Papathanassiou & Geng Wu. 5G Network

Capacity. IEEE Vehicular Technology Magazine. 2014
Moray Rumney. Keysight Technologies - Finding Space for 5G. 2014
Benn, H. Vision and Key Features for 5G Cellular. 2014
http://www.telecomclouds.org/wp-content/uploads/2013/11/. 2015
Afaz Uddin Ahmed, Mohammad Tariqul Islam, and Mahamod Ismail, A
Review on Femtocell and its Diverse Interference Mitigation Techniques in
Heterogeneous Network, Wireless Pers Commun (2014) 78:85106
Zahir, T., Arshad, K., Nakata, A., and Moessner, K. Moessner, K.,
Interference Management in Femtocells, IEEE Communications Surveys &
Tutorials, 15(1):293-311. 2013.
Qualcomm. The Magic of Mobile Broadband: Wireless Fundamentals. 2013
Zahid Ghadialy. 5G: A 2020 Vision. Small Cell SIG & Radio Tech SIG Event,
3rd Feb. 2015
Doug Pulley. LTE Femtocells. 2010

45

2016 Dr.MBI@UKM

Thank you
http://www.ukm.my/mahamod
mahamod@ukm.edu.my
mahamod@gmail.com
019-2615404/019-3275425
03-89216326

Related Research

Problem Statement

Provision of directional beam

forming in femtocell
mandated by coverage
optimization and cell
mitigation
Future 5G wireless networks
will have to contend with
severely limited range at the
high frequencies at which
they will operate
Expect to see a proliferation
of 5G base stations, including
multiple ones within a single
building.
47

2016 Dr.MBI@UKM

Related Research

Problem Statement

A handset usually communicates

though the nearest tower but
can be made to use a more
distant one if the nearest tower
cannot handle its traffic.
No evidence investigating Radio
Environment Map (REM) in
mitigating the intercell
interference.
What is not yet known is the role
of REM in facilitating small and
dense cells deployment in future
5G.
48

2016 Dr.MBI@UKM

Related Research

Problem Statement
Device-to-Device (D2D) architecture improve
throughput, coverage, end-to-end latency.
However, introduces several challenges, such as
interference management between cellular and
D2D users becomes one of the most critical
issues for in-band D2D communication.
If the generated interference is not well
controlled, it will deteriorate the potential
benefits of D2D communication since the overall
cellular capacity and efficiency is degraded

49

2016 Dr.MBI@UKM

Related Research

Objectives
To

introduce a novel DoA estimation technique of

the users in 5G femtocell network by using
machine learning process
To quantify the benefits of REM-data
measurements experimentally in the intercell
interference coordination within 5G small cells
To design an innovative interference cancellation
technique to mitigate cross-layer and co-layer
interference in D2D enabled cellular network.

50

2016 Dr.MBI@UKM

Related Research

Methodology

WP1: DoA Estimation for 5G Femtocell Multi-element Antenna

PHASE 1: Problem background and DoA characterization
PHASE 2: Development of beam steering technique based on
machine learning DoA algorithm
PHASE 3: Validation of beam steering in potential 5G environment
WP2: Interference Mitigation for 5G Small Cells with Radio
Environment Map (REM) PHASE 1: Development of Spectrum Sensing
and Localisation Tracking
PHASE 2: Development of REM database
PHASE 3: Development of Intercell Interference Coordination
technique

51

2016 Dr.MBI@UKM

Related Research
WP 2: Overview of REM Prototype Architecture

52

2016 Dr.MBI@UKM

Related Research

Methodology
WP3:

Cross & Co-Layer Interference Mitigation Strategy

for Device-to-Device (D2D)
PHASE 1: Investigation of interference cancellation
techniques in D2D enabled cellular networks and 5G
transmission
PHASE 2: Exploring the feasibility of integrating
interference cancellation and Beamforming precoding
to D2D enabled cellular network
PHASE 3: Evaluate the interference cancellation based
on 5G specifications and network offloading scenario

53

2016 Dr.MBI@UKM

Related Research
WP 3: Cellular Offloading in D2D
Communications in Multi-tier cells in
Heterogeneous Networks

54

2016 Dr.MBI@UKM