Cognitive Radio For 5G Small - Final Report Done

Department of Electronics
COGNITIVE RADIO FOR 5G SMALL-CELL

COMMUNICATIONS
MSc. in Communications Engineering
Oyomikon Ezekiel Jakpa
Primary Supervisor: Dr. David Grace
Second Supervisor: Dr. Paul Mitchell
10th September, 2014

ABSTRACT ii
ABSTRACT
This project focuses on how to implement a fifth generation (5G) communication system
with the aid of cognitive radio using green base stations. The benefits of cognitive radio
were used to ensure that an energy efficient network was built. A MATLAB simulation
which models such a system is designed with investigation on propagation and loss models;
channel assignment schemes for the access and backhaul networks; and energy saving with
topology management. Focus is on green communications with inspection of efficient
topology management strategies for energy efficiency which is a major issue in modern
telecommunication systems, whilst providing the required quality of service (QoS). The
performance of the system is inspected using signal to noise plus interference ratio (SINR),
blocking and dropping probabilities, and saved energy. The results show that energy savings
of about 60% is achievable at low offered traffic when compared to no topology
management.
O. E. Jakpa, MSc. Project Report, Communications Engineering, The University of York. 2014
CONTENTS iii
CONTENTS
ABSTRACT ................................................................................................................................... ii
CONTENTS ................................................................................................................................. iii
LIST OF ACRONYMS ................................................................................................................... vi
LIST OF FIGURES ....................................................................................................................... vii
LIST OF TABLES .......................................................................................................................... ix
1. INTRODUCTION .................................................................................................................. 1
1.1. Overview ..................................................................................................................... 1
1.2. REPORT OUTLINE......................................................................................................... 3
2. BACKGROUND STUDIES ..................................................................................................... 4
2.1. INTRODUCTION ........................................................................................................... 4
2.2. COGNITIVE RADIO ....................................................................................................... 4
2.3. BEYOND NEXT GENERATION (BuNGee) PROJECT ....................................................... 5
2.4. CONCLUSION ............................................................................................................... 8
3. LITERATURE REVIEW .......................................................................................................... 9
3.1. INTRODUCTION ........................................................................................................... 9
3.2. WHY ENERGY MANAGEMENT? ................................................................................... 9
3.3. GREEN COMMUNICATIONS ...................................................................................... 10
3.4. APPROACHES FOR ENERGY EFFICIENCY .................................................................... 12
3.5. CONCLUSION ............................................................................................................. 14
4. SYSTEM DESIGN AND IMPLEMENTATION ........................................................................ 15
4.1. INTRODUCTION ......................................................................................................... 15
4.2. SYSTEM STRUCTURE .................................................................................................. 15
4.2.1. PHASE 1 .............................................................................................................. 16
4.2.2. PHASE 2 .............................................................................................................. 17
4.2.3. PHASE 3 .............................................................................................................. 17
CONTENTS iv
4.2.4. TRAFFIC AND PROPAGATION MODEL ................................................................ 19
4.2.5. ACCESS NETWORK ............................................................................................. 20
4.2.6. BACKHAUL NETWORK ........................................................................................ 20
4.3. SYSTEM LEVEL ........................................................................................................... 21
4.4. SYSTEM PERFORMANCE MEASURES ......................................................................... 22
4.5. SYSTEM ENERGY MANAGEMENT .............................................................................. 26
4.5.1. ENERGY MANAGEMENT .................................................................................... 26
4.5.2. ENERGY MANAGEMENT RULES ......................................................................... 28
4.5.3. ENERGY EQUATIONS .......................................................................................... 29
4.6. ENERGY MANAGEMENT MODEL (EMM) 101 ............................................................ 32
4.7. ENERGY MANAGEMENT MODEL (EMM) 201 – 202 .................................................. 32
4.7.1. EMM 201 ............................................................................................................ 32
4.7.2. EMM 202 ............................................................................................................ 33
4.8. ENERGY MANAGEMENT MODEL (EMM) 301 – 302 .................................................. 34
4.8.1. EMM 301 ............................................................................................................ 34
4.8.2. EMM 302 ............................................................................................................ 35
4.9. CONCLUSION ............................................................................................................. 38
5. SYSTEM TEST AND ANALYSIS ........................................................................................... 39
5.1. INTRODUCTION ......................................................................................................... 39
5.2. SYSTEM CHECK .......................................................................................................... 39
5.3. TEST AND ANALYSIS IN PHASE 2 ............................................................................... 40
5.3.1. SYSTEM EVALUATION USING PERFORMANCE MEASURES ................................ 40
5.3.2. COMPARISON OF TOPOLOGY MANAGEMENT SCHEMES .................................. 43
5.4. TEST AND ANALYSIS IN PHASE 3 ............................................................................... 47
5.4.1. SYSTEM EVALUATION USING PERFORMANCE MEASURES ................................ 47
5.4.2. COMPARISON OF TOPOLOGY MANAGEMENT SCHEMES .................................. 49
CONTENTS v
5.5. CONCLUSION ............................................................................................................. 50
6. PROJECT MANAGEMENT ................................................................................................. 51
6.1. DELIVERABLES ........................................................................................................... 51
6.2. RISK MANAGEMENT/MITIGATION ............................................................................ 52
6.3. QUALITY ASSURANCE ................................................................................................ 54
7. FURTHER WORK ............................................................................................................... 55
8. CONCLUSION .................................................................................................................... 56
ACKNOWLEDGEMENT .............................................................................................................. 57
REFERENCES ............................................................................................................................. 58
APPENDIX A .............................................................................................................................. 64
APPENDIX B .............................................................................................................................. 65
LIST OF ACRONYMS vi
LIST OF ACRONYMS
5G Virtual Centre of Excellence

ABS Access base stations
BS Base station
BuNGee Beyond Next Generation
CABS Constant ABS
DL Downlink
DoW Description of work
FCC Federal Communications Commission
FSPL Free Space Path Loss
HBS Hub base station
HetNet Heterogeneous network
HSS Hub subscriber station
IT Information Technology
ITU-R International Telecommunications Union -
Radio
LTE Long Term Evolution
MAC media access control
MIMO multiple-input and multiple-output
MS Mobile Station
OFDM orthogonal frequency division multiplexing
OPEX operating expenditure
PHY physical layer
QoS Quality of Service
Rel-8 Release 8
RRM Radio Resource Management
SABS Switchable ABS
SINR Signal to noise plus interference ratio
VCE Virtual Centre of Excellence
LIST OF FIGURES vii
LIST OF FIGURES
Figure 2-1 Example of Utilization of Spectrum holes, directly reproduced from [8] ............... 5
Figure 2-2 BuNGee System Architecture, Directly reproduced from [10] ................................ 7
Figure 2-3 Topology of BuNGee Architecture, Directly reproduced from [14] ........................ 8
Figure 3-1 Power consumption of a wireless cellular network, directly reproduced from [20]
.................................................................................................................................................. 10
Figure 4-1 A simple description of the whole system showing various network segments .. 16
Figure 4-2 Architecture for Phase 1 based on the concept from BuNGee Architecture [14] 16
Figure 4-3 Architecture for Phase 2 based on the concept from BuNGee Architecture ........ 17
Figure 4-4 Architecture for Phase 3 based on BuNGee Architecture ..................................... 18
Figure 4-5 System flow chart .................................................................................................. 24
Figure 4-6 Distribution of users accessing the network for a whole day ............................... 27
Figure 4-7: Energy management flow chart ............................................................................ 31
Figure 4-8 Grouping ABSs based on levels .............................................................................. 33
Figure 4-9 Representation of levels approach with dependencies ........................................ 33
Figure 4-10 Description of the relationship between CABSs and SABSs ................................ 34
Figure 4-11: Topology of level ABS scheme used in Phase 3 ................................................... 35
Figure 4-12: Topology of CABSs and SABSs in Phase 3 ........................................................... 36
Figure 5-1: Number of ABSs in use against Time for a whole day .......................................... 39
Figure 5-2: Number of calls carried by the respective ABSs in phase 3 ................................. 40
Figure 5-3: Percentage Energy Saved using EMM 201 against Offered Traffic (Constant &
Dynamic) for different SINR threshold values ......................................................................... 41
Figure 5-4: Blocking Probability against Offered Traffic (Constant and dynamic) with and
without topology management (EMM 201) ............................................................................ 42
Figure 5-5: Dropping Probability against Offered Traffic (Constant and dynamic) with and
without topology management (EMM 201) ............................................................................ 43
Figure 5-6: Blocking Probability against Constant Offered Traffic for different EMMs and no
TM ............................................................................................................................................ 44
Figure 5-7: Dropping probability against Constant Offered Traffic for different EMMs and no
TM ............................................................................................................................................ 44
LIST OF FIGURES viii
Figure 5-8: Blocking Probability against Dynamic Offered Traffic for different EMMs and no
TM ............................................................................................................................................ 45
Figure 5-9: Dropping Probability against Dynamic Offered Traffic for different EMMs and no
TM ............................................................................................................................................ 45
Figure 5-10: Percentage Energy Saved against Constant Offered Traffic for different EMMs
and no TM ................................................................................................................................ 46
Figure 5-11: Percentage Energy Saved against Dynamic Offered Traffic for different EMMs
and no TM ................................................................................................................................ 47
Figure 5-12: Blocking Probability against Dynamic Offered Traffic for different thresholds
using EMM 301 ........................................................................................................................ 48
Figure 5-13: Dropping Probability against Dynamic Offered Traffic for different thresholds
using EMM 301 ........................................................................................................................ 48
Figure 5-14: Percentage Energy Saved against Dynamic Offered Traffic for different
thresholds using EMM 301 ...................................................................................................... 49
Figure 5-15: Percentage Energy Saved against Dynamic Offered Traffic for EMM 301 &302
.................................................................................................................................................. 50
LIST OF TABLES ix
LIST OF TABLES
Table 1: Division of tasks among team members directly reproduced from the second stage
report. ........................................................................................................................................ 2
Table 2 Comparison of various generations of mobile technologies directly produced from
[12] ............................................................................................................................................. 6
Table 3 : Major Parameters for all Phases [1] ......................................................................... 37
Table 4: Table of deliverables; directly produced but updated from the second stage report
.................................................................................................................................................. 51
Table 5: Risk management ...................................................................................................... 52
INTRODUCTION 1
1. INTRODUCTION
1.1. Overview
The use of radio systems globally is enormous and growing very fast. New services and
technologies are introduced, thereby generating unparalleled demand for radio frequencies
[2]. Worldwide, traffic on mobile broadband systems has grown so fast that it currently
surpasses predictions for 2020 according to Huawei [3].
Despite this rapid growth, radio spectrum is limited; but must be able to handle phone calls
and data traffic with an acceptable level of service [4]. Even though the spectrum can’t be
seen or felt, wireless communication is impossible without it. Its value cannot be
overemphasized; in economic terms, it is now worth over £50 billion a year to the UK [5].
Another area of concern in modern communication systems is energy efficiency. As

different methods are worked on to harness the available spectrum, the deployment of
base stations to meet both capacity and coverage needs has led to an overwhelming
increase in the amount of energy consumed. In the Beyond Next Generation (BuNGee)
Project for example, capacity density of 1Gbps/Km2 is considered for fifth generation (5G)
systems, but with high energy consumption [1]. Therefore, there is a need to consider ways
of meeting capacity needs by maximising spectrum usage in an energy efficient way.
This project investigates the use of cognitive radio (CR) as an energy efficient way of
exploiting the limited radio spectrum. It involves the design and software implementation of
a small cell system based on the cognitive radio concept. The advantages of CR are
harnessed through deployment of base stations in small cells using dual hop architecture to
provide energy saving benefits with coverage and capacity requirements of 5G network.
In this project, the design and implementation was divided it into three phases with
different level of complexity using the concept proposed in the BuNGee architecture. Details
on these phases are provided in chapter four.
This project was carried out by a team of four, with each team member assigned a specific
task. Each member was assigned a technical and administrative role as shown in table 1.
INTRODUCTION 2
Table 1: Division of tasks among team members directly reproduced from the second stage report.
Name Technical role Administrative Duties

role
Oyomikon, -System level and energy Project Chair -Organizes meetings.

Ezekiel Jakpa management
implementation. -Chairs meetings.
Ao, Lei -Traffic model and Deputy Chair -Stands in for the chair.
Propagation and loss
model implementation.
Aisha, -Spectrum sensing and Document -Responsible for

Maukembayeva channel assignment for Coordinator document and code
Backhaul network links version standardization
and transfer.
Ogheneovie, -Spectrum sensing and Project -Takes minutes during

Ajemuta channel assignment for Secretary meetings.
Access network links
-Drafts weekly project
reports.
This report gives details on the aspect I was responsible for. I have worked on energy
management and system level in this project. The energy management aspect was included
since this is vital for present and future telecommunication systems. I have researched some
of the schemes proposed in literature to achieve energy efficiency and implemented sleep
mode as a viable means to reach this goal. At the system level, I have worked with other
team members to get the various aspects of this project together and also evaluated the
performance of the system.
INTRODUCTION 3
1.2. REPORT OUTLINE
Chapter one gives a general overview on the motivation of this project. It briefly highlights
what some researchers have done and the aim of this project.
In chapter two, background information on the some of the key words in this project from
studies carried out is provided. Relevant materials on cognitive radio and BuNGee are
presented.
Chapter three focuses on the review of literature on the subject specific content of this
report. Different ways of reducing the energy consumed were reviewed, including topology
management with the use of sleep modes which have been implemented in this project.
Chapter four provides a detailed description of the system design and implementation at
the various phases. An overview of aspects in this project handled by other team members
is given with details on how energy management is implemented.
Chapter five presents system test and analysis carried out to ascertain the performance of
the system.
Chapter six brings to view how the project has been managed from start to finish. The
challenges encountered while working on the project and how the predicted risk factors and
mitigating circumstance were of benefit are described.
Further work which can be done in the future to meet the aim of this project was
highlighted in chapter seven with a conclusion in chapter eight.
References are presented followed by an appendix showing the codes which have been
used to simulate this project on MATLAB.
BACKGROUND STUDIES 4
2. BACKGROUND STUDIES
2.1. INTRODUCTION
This chapter gives relevant information on the key words in this project. The opening aspect
builds on the overview from chapter 1 to give a better understanding of the functions of
cognitive radio and a summary on the BuNGee project.
2.2. COGNITIVE RADIO
Federal Communications Commission (FCC) defines cognitive radio in [6] as: “a radio that
can change its transmitter parameters based on interaction with the environment in which it
operates. This interaction may involve active negotiation or communications with other
spectrum users and/or passive sensing and decision making within the radio.” Cognitive
radio is seen as the major tool in maximizing the current spectrum band. Haykin [7] pointed
out that if we were to scan the radio spectrum, we would discover that most of the
frequency bands are largely unused which led researchers to think of spectrum holes
defined in [7] as: “a band of frequencies assigned to a primary user (licensed user), but, at a
particular time and specific geographic location, the band is not being utilized by that user.”
This is represented in figure 2.1 where dynamic spectrum access is employed to harness the
unused spectrum.
The main functions of the cognitive radio are [7]:
 Detection of spectrum holes.

 It predicts the channel’s capacity to be used by the transmitter.
 Dynamic spectrum management.
In this project, we have assumed perfect spectrum sensing. This implies that we already
have the necessary information on the available spectrum and also do not need to sense the
spectrum for interference.
Figure 2-1 Example of Utilization of Spectrum holes, directly reproduced from [8]
Mitola [9] proposed the use of cognitive radio to solve current spectrum challenges by
“enhancing the already emerging software defined radio through a radio knowledge
representation language. This language represents knowledge of radio etiquette, devices,
software modules, propagation, networks, user needs and application scenarios in a way
that supports automated reasoning about the needs of the user.”
Although cognitive radio is still being developed, it will give way for effective use of the
spectrum. The transceiver in a cognitive radio scans for unused bands and changes its
transmission and reception parameters to different frequencies during heavy data loads
without interruption. It can also listen for interference on busy channels and calculate a way
to reduce it so that the channels can accommodate more people [4]. In a similar way
cognition has been implemented in this project on the access and backhaul network using
logic. Study was also done on the use of reinforcement learning, but this was not
implemented due to limited time. Also cognition was used to ensure that the base stations
are aware of the state of their environment via its neighbours and adjust its parameters to
meet load demands.
2.3. BEYOND NEXT GENERATION (BuNGee) PROJECT
In a bid to meet the requirements of future demand which has been predicted to occur in
telecommunication systems BuNGee has come up with a novel approach and some
improvements on the already existing technology. This is in order to surpass ITU-R’s
(International Telecommunications Union - Radio) target for compliance to provide

approximately 100-300 Mbps/km2 which will not be sufficient to meet future demands [10].
In urban areas where the population is expected to be dense with more users per square
kilometre, the capacity set by ITU-R will be inadequate. This is seen from a simple
calculation performed in [11] and the description of work (DoW) in BuNGee shows that
around 1Gbps/km2 is needed to meet this demand [10]. Predictions in [12] reproduced in
table 2 also confirm this.
Table 2 Comparison of various generations of mobile technologies directly produced from [12]
The BuNGee architecture has been designed to efficiently meet the rise in demand on
future networks in terms of cost, spectrum and energy. To achieve this, it uses a system
which incorporates the use of Hub base station (HBS), Access base station (ABS), Relay
Stations (RS), Femto-cells and Mobile subscribers (MS) to provide better coverage and a
high capacity data network. For this reason, small cell communication is used in this project
to provide the required network density. Figure 2.2 gives an overview of the BuNGee
architecture showing how the different components of the network are interconnected.
Figure 2-2 BuNGee System Architecture, Directly reproduced from [10]
The BuNGee architecture shown in figure 2.3 is based on a Manhattan grid environment
designed to give good positioning of the network devices as obtained in the real life
situation. Each ABS employs directionality in its antennas pointing up and down the streets
to increase the coverage area and to reduce interference due to the gain of the antenna. At
street intersections, ABS with four directional antennas pointing in different directions is
used to provide same effect in the two direction case. Each ABS has a substation known as
the Hub Subscriber Station (HSS) which is used to connect to the HBS with a backhaul link
[13] [14].
Taking clue from the topology of the BuNGee architecture, similar topologies were designed
for the various phases of this project. This method is better than the hexagonal cell
structure because it is built on already existing structures in the environment. We have not
implemented directionality on the ABSs, but there is directionality on the backhaul links
between the HSSs and the HBSs.
Figure 2-3 Topology of BuNGee Architecture, Directly reproduced from [14]
2.4. CONCLUSION
As discussed, both cognition and the techniques used in the BuNGee project forms the
underlying approach to this project. Some of the concepts used in the BuNGee project were
borrowed but simplified in such a way that we can still model the proposed system.
LITERATURE REVIEW 9
3. LITERATURE REVIEW
3.1. INTRODUCTION
This section provides relevant information on the focus of this report. The need for energy
management and several ways in which it can be accomplished are covered. Using the
findings from the literature review, suggestions were made on how to implement energy
management in this project.
3.2. WHY ENERGY MANAGEMENT?
Given the contribution of subscriber growth, migration to higher rate broadband and that of
IT (Information Technology) to global energy consumption, it is necessary to reduce the
amount of energy required. Excluding the power consumed by mobile phones, its
supporting network in the United Kingdom consumes about 40–50 MW [15]. Vodafone
expends over 1 million gallons of diesel every day to sustain their network in under
developed countries. This makes it obvious that mobile communications sector contributes
significantly to energy consumption worldwide [15].
The essential cycle of developments in telecommunications which supports economic

growth and as a result support advances in telecommunication has been of great benefit.
Nonetheless, this cycle will not last without end since all telecommunication networks
require resources to function, for the most part electrical power to operate. As the network
continues to grow with tremendous increase in its infrastructure and capacity more electric
power is also consumed. Research done in line with this confirms that as the world economy
becomes more networked with more activities going online, the power consumed will
continue to increase [16]. This is one of the challenges the telecommunication sector is
currently faced with and it will give rise to greater problems in the future if not properly
tackled [17]. In the present day, the information and telecommunication sector is
accountable for about 5% of the overall electrical power consumed in industrialized national
economies [18, 15] and about 2 to 2.5% of the total energy consumed worldwide [19]. In
order to provide solution to this problem we must first be able to give a near accurate
estimate of the power consumed and where most of it is consumed.
LITERATURE REVIEW 10
3.3. GREEN COMMUNICATIONS
The term Green Radio was first used by Mobile VCE's (Virtual Centre of Excellence) in early
2007 and has since become a household name in the telecommunication industry. Owing to
the rapid growth of the telecoms sector, lowering carbon emissions and operating
expenditure (OPEX) costs for wireless networks is essential and they are the two major
factors that propels Mobile VCE’s Green Radio programme to aim at achieving a hundred
times decrease in energy consumption without a trade-off on the quality of service at the
user end [20].
Taking into account the power consumed by the various components listed above, the total
base station power consumption will drastically reduce with implementation of energy
management. From figure 3.1, the chart shows that in a wireless cellular network, most of
the power is consumed by the base stations. Therefore, this project considers different ways
of ensuring efficient energy consumption in the base stations with an acceptable quality of
service.
Cellular Network Power Consumption
Figure 3-1 Power consumption of a wireless cellular network, directly reproduced from [20]
Most of the energy consumed in current and future systems will come from the base
station. Hence, its power consuming components such as the rectifiers, transceiver,
microwave links and cooling system should be taken into consideration when accounting for
the energy consumed by the base station [21].
In modelling this aspect of the telecommunication network, two methods are majorly used.
These are;
 Equipment inventory and sales figures [18, 22] and

 Telecommunication network design principles [23, 24].
The first method which is generally accepted for modelling the system according to
equipment inventory and sales figures depends on the data from previously sold
telecommunications equipment to provide an estimate of the quantity of equipment in the
network and the amount of energy consumed by this equipment.
The advantage of this method is that it helps in estimating ways in which increased internet
activities in the future will affect how power is consumed. Although this is important, it does
not clearly show the relationship between increasing number of users of these equipment
and the consequential power consumption.
Hence the second method which is usually implemented alongside the first is done using the
design principles of the telecommunication network. From the design in [23, 24], it is
possible to estimate the amount of energy consumed in each part of the network using the
actual factory rating of those equipment and their different access rates. This method allows
for a more accurate model in estimating the amount of power consumed and how it will
appear in the future with rise in demand (capacity) due to increased number of users [17].
Energy efficiency is also necessary to handle challenges regarding increase in emission of

CO2, depleting fossil fuel sources and escalating electricity price [20, 25]. This will not only
result in a healthy environment, but will aid efficient management of resources and lower
fares charged by telecommunication service providers.
3.4. APPROACHES FOR ENERGY EFFICIENCY
In order to solve the existing challenges in meeting high capacity and coverage needs, the
design and deployment of infrastructures on the network are being researched and
improved. In [26] deployment schemes are considered to have high impact on energy
consumption. Some approaches considered here are:
 Radio Resource Management (RRM) Strategies: To achieve energy reduction by using

RRM strategies, in Long Term Evolution (LTE) some schemes like the orthogonal
frequency division multiplexing (OFDM) and multiple-input and multiple-output
(MIMO) are used for the downlink (DL) transmission [27]. RRM aims at decreasing
the amount of delay that occurs during transmission of packets, give better
spectrum flexibility, increase the efficient use of the sparse radio spectrum resources
and network infrastructure and cost reduction for operators and end users. They are
majorly used in the physical layer (PHY) and media access control (MAC) layers to
reduce energy consumption in the base station. RRM allows the system on the
network to have control over its fundamental parameters like the power used for
transmission, allocating channels, modulation order and error coding scheme, etc
[28].
 Maximizing data rate: Logically, the power consumed in a network may not be easily
reduced especially at high traffic. Hence, energy efficiency in terms of data rate
refers to maximising the size of data that can be transferred on a network from one
point to another relative to the amount of expended energy. This degree of energy
efficiency is basically a reciprocal of the energy per bit of data transmitted [17]. In
[29] a method for proper allocation of the spectrum is proposed to maximize the
data rate per unit energy.
 Antenna muting: This is a good method that functions on a fairly limited time scale
so as to decrease the energy consumed in an LTE cell [30]. If an antenna has two
ports, one of them may be switched off at off peak periods as proposed in LTE
Release 8 (Rel-8) [31]. This is because during off peak period, the traffic and inter-cell
interference are usually low and the diversity provided by the use of multiple
antennas is not needed. At low offered traffic, the radio units operate below their
full output power; hence it’s possible to increase the power on other radio units. Up
to 50 percent reduction in energy can be achieved at low offered traffic without
considerably altering user throughput [30].
 Sleep mode: Based on the activity of the mobile station on the network, a base
station can be turned off by temporarily shutting down most of its power consuming
components as seen in [32]. Since the base stations will be underutilized at various
time in the day especially when the offered traffic is low, a flexible base station
switching strategy with the use of cooperative communications was proposed in [33]
and the schemes proposed were examined to give up to 50% power saving [33]. In
this project we have majorly implemented sleep mode, by turning off a base station
when the traffic falls below a specified threshold.
 Heterogeneous Network (HetNet): This is a good technique for intensifying
mobile network capacity. A HetNet usually comprise of multiple radio access
technologies (e.g. Bluetooth, Wi-Fi, and 3G, 4G or LTE), architectures, and base
stations which are able to adapt their transmission power [34]. This is one of the
major technologies used at present to reduce energy consumption. In such
networks, the cell size is varied by the use of femtocells to provide increased
capacity and coverage in regions where the signal tends to be poor with a range of
10 to 30dBm and maximum transmit power of 20 to 30dBm [25]. However, it has
some inference issues which must be taken care of even though the network is
deployed in an energy efficient way [35, 36].
 Topology management scheme: The term topology generally means layout and in
this case, topology management refers to the management of the architecture in
which the base stations are located. This differs from one geographical region to
another, since the way the base stations will be located largely depends on the
nature of the environment (population density, network of roads, houses, and other
existing structures). Hence, during the installation of the base stations these factors
could be used to determine which base station remains on all the time and those
that can be turned to sleep mode without going below an acceptable level of service.
The use of macro-cell concept together with the BuNGee novel architecture was
proposed in [37] as a scheme that provides better energy savings than the state-of-
the-art topology management schemes implemented in the BuNGee systems.
3.5. CONCLUSION
Since most of the energy is consumed at the base station [38], in this project different
model of sleep mode mechanisms are used to keep the base station idle when it is not in
use. This is because at some point during the day, the traffic load is usually low making it a
reasonable idea to switch off base stations which are not in use or those operating below
their capacity, bringing about a more energy-efficient network operation [33].
SYSTEM DESIGN AND IMPLEMENTATION 15
4. SYSTEM DESIGN AND IMPLEMENTATION
4.1. INTRODUCTION
This chapter focuses on the entire system architecture which is based on the BuNGee
architecture, using similar principles in the deployment of MSs, ABSs, HSSs and HBSs. An
overview of the concepts employed by other members of the team for optimum operation
of the whole system is also presented as it is necessary to have a good grasp of the system
level operations. Two energy management schemes were considered and performance
measures were suggested.
4.2. SYSTEM STRUCTURE
In our system, the traffic and the associated losses are first generated with the MS static
throughout a call period. To ensure that the system is energy efficient, the traffic is
monitored so that ABSs not in use are put to sleep and based on set rules some may be
switched on where necessary to meet MS demand. The MS uses two hops to get connected
to the HBS as shown in figure 4.1. This implies that MS will first be connected to the ABS and
then to the HBS via the HSS. No MS is able to directly get its service from the HBS. The dual
hop architecture helps to reduce energy consumption and distributes the load on the
network.
HBS
BACKHAUL NETWORK
Second
NETWORK
hop
ABS
ACCESS NETWORK
First
hop
MS
Figure 4-1 A simple description of the whole system showing various network segments
4.2.1. PHASE 1
Designed to get the project started, this phase is kept simple in order to give every team
member a good understanding of the basic concepts used. The topology for phase one is
shown in figure 4.2. The system comprise of three users at off peak and nine users at peak
period, with four ABSs, four HSSs and one HBS deployed in a 300m by 300m area with six
channels on the access network and four channels on the backhaul network.
Legend:
-ABS/HSS
-HBS
Figure 4-2 Architecture for Phase 1 based on the concept from BuNGee Architecture [14]
4.2.2. PHASE 2
The level of complexity in this phase was increased in order to get close to what is
obtainable in real life situations. Here we have used one hundred MSs, evenly distributed
across the 800m by 800m service area. Thirteen ABSs with each of them having one HSS and
two HBSs (one for each cell) are deployed with the same number of channels as in phase 1.
Directionality was also incorporated into the links between the HSSs and the HBSs to
increase the performance of the system.
Legend:
-ABS/HSS
-HBS
Figure 4-3 Architecture for Phase 2 based on the concept from BuNGee Architecture
4.2.3. PHASE 3
This phase of the project has the highest complexity. There are one thousand MSs, seventy
six ABSs each with an HSS and five HBSs deployed in a 1500m by 1500m area, similar to that
in the topology of the BuNGee architecture in figure 2.3.
Legend:
-ABS/HSS
-HBS
Figure 4-4 Architecture for Phase 3 based on BuNGee Architecture
4.2.4. TRAFFIC AND PROPAGATION MODEL
The traffic and propagation model worked on by Ao Lei in this project is fundamental
because no other aspect of the network can be properly designed and tested if there is no
traffic generated with simulation of the environment in which the transmission of the traffic
occurs.
In the traffic model, the call arrival and departure times follow an exponential distribution
and comparison of the result from Monte Carlo and theoretical simulations showed that it
was properly executed. Two traffic models have been implemented in this project; these
include constant and dynamic traffic models.
The constant traffic model was used in phase one and two only to give about an equal
amount of traffic for a whole day, while the dynamic traffic model was used in second and
third phases to simulate varying offered traffic which depends on the different weight of
traffic on the network for a whole day.
In the propagation model, losses in transmission and interference from other antennas were
modelled. We have used free space path loss (FSPL) in the first phase of this project because
of the simple topology and reduced obstruction at that level, which will make it possible for
FSPL to give reasonable results. The equation below is used for simple path loss estimation
[39]:
( ) (( ) ) (1)
Where and d represents wavelength and distance from the transmitter respectively.
Winner II B1 path loss model was used in the second and third phases because it is suitable
for the topology deployed at those levels, in terms of its frequency range and adaptability to
the simulated environment. The Winner II model, path loss is evaluated by the equation
[40]:
[ ]
( [ ]) ( ) (2)
In equation 2; d represents distance between the transmitter and receiver, fc is the

frequency, A is path loss, B is the intercept and C is the frequency dependent parameter.
More information on the traffic and propagation model can be found in Lei’s report.
4.2.5. ACCESS NETWORK
The access network as shown in figure 4.1 consists of the ABSs and MSs. The ABSs use omni-
directional antennas to provide the required network coverage. The access network is under
the first hop, and it considers issues in assigning channels to MSs, while taking into account
the interference coming from other MSs and ABSs on the network.
Different schemes for channel assignment were considered in order to provide an

acceptable level of service, and to make good comparisons. Some of the channel assignment
schemes considered are; best SINR which was implemented in all phases of this project and
fixed channel assignment which was implemented in phase 2. Random channel assignment
and reinforcement learning were also worked on, but they gave poor performance.
Ogheneovie has more information on this aspect presented in his report.
4.2.6. BACKHAUL NETWORK
The backhaul network is also shown in figure 4.1 consisting of the HSSs and HBSs
establishing backhaul links with directionality on their antennas to reduce interference and
provide better performance. The backhaul network is under the second hop and it gives the
already accepted users from the access network admission to the HBS.
The same schemes used on the access network were implemented on the backhaul
network. Detailed information on this can be found in Aisha’s report.
4.3. SYSTEM LEVEL
At the system level, the various aspects of the project are integrated together. Simulation
performed shows the behaviour of users in the system, how it affects topology
management, efficient use of energy, reduced interference, spectrum sensing and channel
assignment.
The operation of the system is highlighted in the eight levels below:
 Generate MS, ABS, and HBS Location: At this point, we generate the location of the
users or MS, the ABS, and the HBS.
 Generate Matrix of Losses: Loss models are set for the access link and the backhaul
link based on the users location.
 Generate Traffic: Using a Poisson distribution, the traffics of the users are generated.
This determines their call arrival time and the time between calls.
 Energy Management: Based on the average traffic level of an ABS and the state of its
neighbours, it switches OFF or remains ON.
 Request for Channel: The incoming user request for available channel in the system.
 Spectrum Sensing: The system checks for available channels for the new user (MS),
and checks if the SINR level is above the set threshold.
 Channel Assignment: If a channel is available with SINR threshold lower than that of
an incoming user, the user is assigned such a channel. There are two channel
assignments done here, one for the access link and another for the backhaul link.
The MS first connects with the ABS and if it is successful, it is then transferred from
the ABS to the HBS via the HSS.
 Session Recorded: The arrival of users is recorded, and when the channel is full, no
other user is allowed in.
We are able to check the performance of our system via some measures set in the
algorithm.
4.4. SYSTEM PERFORMANCE MEASURES
Energy efficient algorithms are said to be effective when it is not only able to save energy
but also sustain the quality of service on the system [41]. Therefore, the performance of the
system designed in this project is evaluated using some of the standard measures in a
telecommunication system. Performance was measured using signal to interference plus
noise ratio (SINR), blocking and dropping probabilities and saved energy, with comparison
on theoretical values and other results proven from some published and accepted research.
 Signal to Interference plus Noise Ratio: In a wireless network with m transmitter or

receiver pairs as designed in this project, when a particular transmitter is
communicating with a receiver, other transmitters act as source of interference to
the main transmitter. This is severe when there is no directionality on the antenna as
used in the access network compared to backhaul where we have implemented
directionality. The receiving nodes also have its noise power which we have taken
into consideration. In equation 3;
o Si is the desired signal power from transmitter i,
o Ii is the total noise power received from all the other receivers and
o i is a measure of the receiver noise.
The SINR level at each receiver is measured using the equation [42]:
( ) (3)
 Blocking probability: This is the probability that a connection request will be blocked
[43]. Cellular systems are designed to a blocking rate of 1% to 5% at high offered
traffic, with 2% being the most common. The blocking rate is usually lower when the
offered traffic is low [44]. Blocking occurs in our system when no channel is available
for the incoming call or the SINR of the incoming call does not meet quality of service
requirements. Where PB is the block probability, NB is the total number of dropped
calls and NB is the total number of accepted calls, equation 4 gives a mathematical
description of blocking probability on a network [14]:
(4)
 Dropping probability: This refers to the phenomenon of call dropping in voice

networks. Call dropping refers to the termination of a call before its intended end
[45]. In our system this occurs when the channel assigned to a user is no longer
suitable for the ongoing call because the SINR of the user falls below the system
threshold. Where PD is the drop probability, ND is the total number of dropped calls
and NA is the total number of accepted calls, equation 5 gives a mathematical
description of dropping probability on the network [14]:
(5)
 Saved Energy: As a performance measure, this is used to analyze the amount of

energy reduction achieved when the different energy management models are
implemented on the network. To do this, the total energy consumed when topology
management (ETM) scheme is used, is compared with the total energy consumed
without topology management (ET). Below is a simple equation used to compute the
percentage of saved energy (%ES) [14]:
( ) (6)
The system flow chart in figure 4.5 gives a clear description of the algorithm with which the
overall system works.
Figure 4-5 System flow chart
At the start of the simulation, the propagation model is initialized to generate the location
of the MSs and Base Stations (BSs). The BSs are fixed according to the set topology, while
the MSs are uniformly distributed across the cells. From the location of MSs, their distances
are calculated and used to generate a loss matrix based on the propagation model in use.
Next, the traffic model allows for call arrival and departure times are generated for all the
MSs.
If there is call departure, the previously occupied channel is released and energy
management is carried out by switching off the appropriate BS according to the specified
rules. In the case of an arrival, energy management is still carried out, by deciding which BS
to switch on, in order to meet an acceptable level of service.
Afterwards, the MS makes request for a channel on the access network. The MS is not
allowed into the system if its SINR value does not meet the required threshold based on the
schemes implemented on the access network. This is referred to as a blocked call.
Otherwise, the call is accepted and a check is carried out on the ongoing calls, if they meet
the acceptable level of service. A call is dropped if it is termed not to be ok based on the
scheme implemented, for example best SINR.
The access network has been directly integrated with the backhaul network; so the calls
received from the access network are recorded as accepted calls to be assigned a channel
on the backhaul network. The same process of channel assignment on the access network is
carried out on the backhaul network for users already in the system. In this case, when a call
is blocked or dropped, its corresponding channel on the access network is released.
The iteration continues until the end of events and the results are analyzed to check the
performance of the system.
4.5. SYSTEM ENERGY MANAGEMENT
Current research for future systems propose the use of multiple hop networks to provide
sufficiently high throughput densities with better coverage. For this to be possible, more
base stations are needed within the service area. At present, such base stations are built to
deliver very good performance without energy efficiency and results 90% of the total energy
consumed [46]. With the current design techniques for future networks, the increasing
number of base stations deployed will result in a tremendous increase in the energy
consumed [47]. When the traffic load is low, most of these base stations will operate below
their capacity while maintain the energy cost of running them.
To solve this problem, this project considers the energies associated with the base station
and attempts to locate where and how substantial level of energy can be saved. By taking
proper account of the energy expended by the base station, it has been noted that there are
two major types of energy which are associated with the base station, these include [48]:
 Embodied Energy: This refers to energy consumed in the extraction of unrefined

resources, its transportation, production, assembly, setting up of the base station
and its services, including disassembly, deconstruction and decomposition.
 Operational Energy: Energy consumed throughout the operational lifetime of the

base station.
The operational energy usually predominates over embodied energy owing to the lengthy
lifespan of the base station which is about a period of ten to fifteen years [48]. Hence, only
operational energy in considered in this project.
4.5.1. ENERGY MANAGEMENT
In order to achieve significant reduction in the energy consumed, networks must be

designed to be energy efficient by deploying sleep modes. Sleep mode is said to be in use
when wireless resources and hardware devices which are not in use are switched off,
resulting in some saved energy which would have been expended while the system was not
used. At present, sleep modes are extensively applied to several wireless devices to help
conserve the lifespan of batteries [49].
Sleep mode was implemented in this project as an efficient way of managing the network
topology because of the large number of base stations in use. This is a reasonable option
since the plot of the distribution of users accessing the network shown in figure 4.6, gives us
a clue about the behaviour of the users throughout a 24 hour period. The number of users
accessing the network is lowest around 2:00 to 4:00 hours in the morning and peaks at
about 20:00 to 22:00 hours in the night. Therefore, it is not an efficient technique to keep all
the ABSs on throughout the day. Some rules introduced, tested and modified in [47, 50, 51]
will be used to switch off (put to sleep) some ABSs based on traffic, to save energy on the
overall network. A situation in which the base station is kept on throughout the day without
any form of energy management is referred to as no topology management (No TM).
1400
1200
1000
Number of Users
800
600
400
200
0
2 4 6 8 10 12 14 16 18 20 22 24
Time (hrs)
Figure 4-6 Distribution of users accessing the network for a whole day
Energy management in this project aims at minimizing energy consumption whilst providing
an acceptable quality of service. To achieve this, ABSs and HBSs which are not in use is kept
in sleep mode. The term, “switching OFF the ABS” has also been used interchangeably with
sleep mode. When in sleep mode, the ABSs uses pilot signals to relay state information of its
environment to the HBS as a way of cognition and saving more energy.
4.5.2. ENERGY MANAGEMENT RULES
In order to efficiently implement energy management, the topology is adjusted to ensure

that all network connectivity parameters are properly managed and that its values are
within certain bounds [52].
In this project, the ABSs have been designed to intelligently sense their environment, and be
in either an on (active) or off (sleep) state based on its average usage (capacity) and the
state of its neighbours. The general rules for switching on and off the ABSs are stated below
[50, 51]:
Switch OFF rules for ABS:
 If the ABS average load capacity is lesser than 50% for Ts and
 All ABS neighbour load capacities is lesser than 25%
Switch ON rules for ABS:
 ABS neighbour load capacity greater than or equal to 50% for Ts and
 If blocking occurs
ABSs can be switched ON if neighbours become overloaded or blocking exceeds the

threshold in a zone. Where Ts represents 10 seconds waiting time [51].
The average capacities of ABSs, its neighbours and Ts were chosen to prevent unnecessary
oscillations between the on and off modes. For example, without specifying the waiting
time before the base station is put to sleep, it will continue to alternate between the on and
off modes as users arrive or departs from the network, leading to more energy consumption
and reduction in the life span of the base station.
4.5.3. ENERGY EQUATIONS
With our focus on uplink since its impact is similar to that of the downlink, an understanding
of the components resulting in most of the energy consumption from [26], and operational
energy according to [24], the energy expended in the MSs, ABSs and HBSs when sleep mode
is implemented will be estimated with the aid of mathematical equations. The energy
consumed by the MS during transmission is given by the equation [47]:
∑ (7)
In equation 7 , nms, tMS,i and PMS,Tx,i represents the total number of MSs in system, its service
time and the transmit power.
In the ABS, the total energy consumed is based on the different states it moves to from time
to time, which is determined by the traffic model. This includes the idle, receive, active, and
transmit modes. The energy consumed by the base stations are shown in the equations 8
and 9 [47]:
∑( )( )
∑ (
(8)
)( )
where t, P,n,E represents time duration, power, total number and energy respectively. The
subscripts define how these variables have been used.
The energy consumed in the HBS is calculated by summing the total time in receive mode,
power consumed in receive and active modes for a specified number of HBSs.
∑ ( )( ) (9)
Details of equations 7, 8 and 9 are given by Han in [47].
Another aspect of energy management on the network is to maintain a certain kind of

topology such as a tree topology [52]. In which case, a unique approach has been used in
this project to maintain a unique topology based on levels. An approach proposed in [53]
where one ABS controls three other ABSs, switching them on and off with respect to their
average capacity is also used to give good comparison. The flow chart below describes the
operation of the energy management models used.
Traffic
Traffic
Switch off Analyse switch Switch on

conditions
Satisfy switch Satisfy switch

No No
off conditions? on conditions?
Yes
Yes
Switch off Switch on

ABS ABS
Record off status Record on status
and duration and duration
Calculate the total

energy consumed
End
Figure 4-7: Energy management flow chart
Based on the traffic on the system, a decision is made whether it is necessary to switch an
ABS on or off according to the energy management rules. If the switch off rules is satisfied,
the ABS concerned is switched off and a count is done in order to keep track of the total
number of off times and the duration. If not, it is noted that the base station is still on. In
the case where the switch on was satisfied, the corresponding ABS is turned on, with its
status recorded else it implies that the ABS is still off.
The on and off duration of the ABSs and the number of time it moved from off to on state
are applied in the energy equation to calculate the total energy consumed.
In the models described below, 101 means “first phase first scheme,” 201 means “second
phase first scheme,” 202 means “second phase second scheme,” 301 means “third phase
first scheme” and 302 means “third phase second scheme.”
4.6. ENERGY MANAGEMENT MODEL (EMM) 101
Phase 1 energy management was carried out to get the project started. Since there are four
ABSs in this phase, two opposite ABSs are linked up so that they either switch off or on at
the same time in order to maintain good network coverage. The idea is to switch off an ABS
in a situation where it’s capacity is less than 50 percent or its not in use for at least 10
seconds and the capacity of its neighbour is less than 25 percent. When the ABSs currently
in use have a capacity greater than or equal to 50 percent, with at least one blocked call, the
neighbour ABSs are switched on.
4.7. ENERGY MANAGEMENT MODEL (EMM) 201 – 202
4.7.1. EMM 201
EMM 201 is unique approach implemented in this project to maintain good topology of the
network as ABSs move from one state to another (i.e. on or off state). In addition to the
energy management rules, this model turns off the ABSs based on levels. This is done
because from the observation of traffic load on each base station for a complete session,
the ABSs at the edge of the network are seldom used. They are therefore used as level 3
ABSs while the next most frequently used ABSs are set as Level 2 ABSs. Level 1 ABSs which
are always kept on. Figure 4.8 shows the ABSs identified by numbers and arranged in
hierarchy. The arrows used in figure 4.9 shows dependencies. It describes which ABS is seen
as a neighbour by another ABS.
Figure 4-8 Grouping ABSs based on levels
1 2 3 4 5
6 Legend:
13 10
Level 1 ABS:
Level 2 ABS:
Level 3 ABS:
12 11 9 8 7
Figure 4-9 Representation of levels approach with dependencies
4.7.2. EMM 202
To achieve good comparison of different schemes, the EMM 202 has also been
implemented in this project, based on a novel approach introduced in [53]. Instead of
switching by levels, a concept known as Constant ABSs (CABS) and Switchable ABSs (SABS)
was used. The CABSs were carefully selected from the system after observing the ABSs that
were frequently used and would give the required coverage in the area where it is located in
the absence of other ABSs. The SABSs are selected from those ABSs not frequently used and
whose absence can be well catered for by CABSs especially at low offered traffic. The CABSs
are on throughout the day, while the SABSs are switched on when the traffic level increases
in order to meet the capacity demand.
In this phase, the four ABSs at edge of the network are used as SABSs together with four
other ABSs based on their average usage whilst ensuring that the CABSs chosen can provide
acceptable coverage.
Legend:
CABS:
SABS:
Figure 4-10 Description of the relationship between CABSs and SABSs
4.8. ENERGY MANAGEMENT MODEL (EMM) 301 – 302
4.8.1. EMM 301
This model is the same as that implemented in EMM 201, with upgrade and modification
due to the change in the system topology and the number of ABSs used in phase 3. The
same concept of levels is used, with each ABS having dependencies which decides the ABSs
it can communicate with as neighbours. In this case none of the ABSs is kept constantly on
because of the even distribution of the ABSs.
1 2 3 4 5 6 7 8 9 10
11 12 13 14
15 16 17 18
19 20 21 22
23 35
24 25 26 27 28 29 30 31 32 33 34
36 37
38 39
40 41
43 44 45 46 47 48 49 50 51 52 53
42 54 Legend
55 56 57 58 Level 1
Level 2
59 60 61 62
Level 3
Level 4
63 64 65 66
67
68 69 70 71 72 73 74 75 76
Figure 4-11: Topology of level ABS scheme used in Phase 3
4.8.2. EMM 302
This model is about the same with that implemented in [53] because the second stage of
that project is quite similar to the third phase here. The model introduced in that project
uses the concept of CABSs and SABSs to reduce energy consumption while offering an
acceptable QoS. The CABSs and SABSs were chosen in the same way described in EMM 202.
Each CABS controls three other switchable SABSs and its switches them on and off based on
the capacity. Figure 4.12 show this topology, with the CABSs colored half way and located at
the center of each cell.
1 2 3 4 5 6 7 8 9 10
11 12 13 14
15 16 17 18
19 20 21 22
23 35
24 25 26 27 28 29 30 31 32 33 34
36 37
38 39
40 41
43 44 45 46 47 48 49 50 51 52 53
42 54
55 56 57 58
59 60 61 62
63 64 65 66
67
68 69 70 71 72 73 74 75 76
Figure 4-12: Topology of CABSs and SABSs in Phase 3
The table of some parameters used in the project as updated from phase 2 in shown in table
2. This table has been made available to every member of the group after it was updated by
Oyomikon E. Jakpa.
Table 3 : Major Parameters for all Phases [1]
Parameters Values
Phase 1 2 3
HBS 1 2 5
ABS/HSS 4 13 76
Users Off-Peak: 3 100 1000

Peak: 9
Dimension 300m*300m 800m*800m 1500m*1500m
HBS antenna gain 13.2dB 13.2dB 13.2dB

omni-
directional
HBS height 25m
ABS/HSS height 5m
MS height 1.5m
HSS 13.2dB
ABS antenna gain 12.5dB omni-directional
HBS transmitted power 38dBm
ABS transmitted power 38dBm
HSS transmitted power 38dBm
Carrier frequency 3.5GHz
MS transmitted power -7dBW
Bandwidth(what is the impact on the work) 10MHz
Number of Access network channels 6
Number of Backhaul network channels 4
SINR Threshold 1.8dB
Temperature 295K
Boltzmann constant -229dBW/K/Hz
ABS Operating Power 5W
Sleeping Coefficient of ABS 5% of ABS operating power (250mW)
HBS Operating Power 60W
4.9. CONCLUSION
From concepts used in the BuNGee architecture, typical topologies were suggested for the
various phases of this project to get the work started and emulate real life scenarios. It is
important to note that directionality has been deployed on the backhaul network, to give
improvement in capacity due to higher frequency reuse. Dual hop architecture was also
used to allow for reduction in the transmit power used in base stations. Some energy
management models were proposed to aid efficient topology management with an
acceptable performance on the overall system.
SYSTEM TEST AND ANALYSIS 39
5. SYSTEM TEST AND ANALYSIS
5.1. INTRODUCTION
The results gotten based on the set performance measures and suggested energy
management models are analysed in this section. These include the SINR levels, blocking
and dropping probabilities, and energy saved with the proposed EMM schemes. I have
plotted some of the graphs here with Ao Lei and made them available to other team
members where necessary. Therefore similar plots may appear in other reports submitted
for this project.
5.2. SYSTEM CHECK
The plot in figure 5.1 clearly shows the average ABS usage for a whole day when the offered
traffic in phase 3 is dynamic. At the early hours of the morning when traffic is low, there are
relatively few ABSs in use until about 5:00 hours. Around this time there is an increase in
offered traffic, leading to a corresponding increase in the number of ABSs in use. Peak
period occurs at about 22:00 hours where there is highest traffic and number of active ABSs.
This confirms the realization in chapter4 on the need to switch off some ABSs which will not
be in use at certain times.
70
65
60
Number of ABSs in use
55
50
45
40
35
30
25
0 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
Time (hour)
Figure 5-1: Number of ABSs in use against Time for a whole day
Figure 5.2 shows the amount of calls carried by each ABS for a whole day and we can see
that the number of calls that goes through the ABSs varies according to its position and the
weight of traffic. Some ABSs are observed not to carry a lot of traffic and this could be
likened to sparsely populated areas in a real life scenario where only few users are mostly
available. This further buttresses the need to implement energy management on
telecommunication systems.
4
x 10
4
3.5
3
Number of Calls
2.5
1.5
0.5
0
0 10 20 30 40 50 60 70 80
ABS Number
Figure 5-2: Number of calls carried by the respective ABSs in phase 3
The first phase of this project was done to get the project started, so the performance
measures were not fully considered. Although energy savings of about 40 to 60 percent was
recorded with zero blocking and dropping probabilities, this phase was not designed to
simulate real life scenarios. Hence, it is not analysed here.
The performance measures described are for the whole system with specific test carried out
on the two schemes implemented.
5.3. TEST AND ANALYSIS IN PHASE 2
5.3.1. SYSTEM EVALUATION USING PERFORMANCE MEASURES
Figure 5.3, the plot shows the percentage of energy saved for both constant and dynamic
offered traffic using different acceptance SINR threshold levels. Generally, more energy is
saved at low offered traffic for both cases (constant and dynamic) because there are few
users on the system and most of the base stations are put in sleep mode since they will not
be in use.
Nevertheless, as the offered traffic increases there will be degradation of service. To combat
this, more base stations will need to be turned on and this will lead to higher energy
consumption and reduction in saved energy.
Acceptance threshold of 1.8dB gives a slightly better performance in terms of energy saved
than 5.4dB because only users with SINR lower than the specified threshold are granted
access into the network. This prevents more base stations from turning ON, thus saving
energy.
About 60% energy is saved at low traffic load when compared with the no topology
management scheme. At high traffic load, about 35% energy savings is still achieved for
dynamic offered traffic and about 5% for constant offered traffic.
70%
Constant Traffic - Threshold 1.8dB
Constant Traffic - Threshold 5.4dB
60% Dynamic Traffic - Threshold 1.8dB
Dynamic Traffic - Threshold 5.4dB
Percentage Energy Saved
50%
40%
30%
20%
10%
0%
5 10 15 20 25 30
Offered Traffic (Erlangs)
Figure 5-3: Percentage Energy Saved using EMM 201 against Offered Traffic (Constant & Dynamic) for different SINR
threshold values
In terms of blocking, dynamic offered traffic without topology management (energy saving)
gives the best performance between 0 and 4.5%. This is expected because all base stations
remain ON throughout the day and only interference or capacity limitations will result in
blocking which seldom occurs due to the system design. For constant offered traffic, the
blocking probability is higher without topology management because of the constant level
of offered traffic. The deviation from about 20 Erlangs is most likely because of the high
traffic load which exceeds the capacity of the system. It is also possible that at those
periods, most users do not meet the set threshold.
With topology management (energy saving), both constant and dynamic traffic have higher
probabilities at about 9.5% and 7% respectively. This increase is due to blocking that will
arise because some base stations are in sleep mode and are unable to serve at that instant.
This is however a good trade-off for the saved energy since the blocked user is likely to
again access on the next trial.
0.1
0.09 Constant Traffic - With Energy Saving

Constant Traffic - Without Energy Saving
0.08 Dynamic Traffic - With Energy Saving
Dynamic Traffic - Without Energy Saving
0.07
Blocking Probability
0.06
0.05
0.04
0.03
0.02
0.01
0
5 10 15 20 25 30
Figure 5-4: Blocking Probability against Offered Traffic (Constant and dynamic) with and without topology management
(EMM 201)
As expected, the dropping probability is generally lower than the blocking probability
because users with SINR values below the set threshold have been dropped and only those
that can be offered an acceptable QoS are considered for dropping. On the access and
backhaul networks, another calculation is done after admittance of a user to check the new
SINR value on an ongoing call. However, it will also be dropped in situations where it falls
below the threshold.
There is a corresponding increase in the dropping probability as offered traffic increases.

The topology management and the offered traffic (constant or dynamic) in the system also
affected the performance in the same way as explained for blocking probability.
Although the system gives a better performance without energy saving, there is also a good
trade-off between energy saved and the dropping probability. When topology management
is used there is 8% and 4% dropping for constant and dynamic traffic respectively, and the
system can be said to have a good performance.
0.09
Constant Traffic - With Energy Saving

0.08 Constant Traffic - Without Energy Saving
Dynamic Traffic - With Energy Saving
0.07 Dynamic Traffic - Without Energy Saving
0.06
Dropping Probability
0.05
0.04
0.03
0.02
0.01
0
5 10 15 20 25 30
Figure 5-5: Dropping Probability against Offered Traffic (Constant and dynamic) with and without topology
management (EMM 201)
5.3.2. COMPARISON OF TOPOLOGY MANAGEMENT SCHEMES
In figure 5.6 the blocking probability when EMM 201 is used is compared with that of EMM
202 and no topology management scheme. EMM 202 is observed to give a slightly better
performance at high level of offered traffic than EMM 201. At low level load EMM 202 gives
a far better performance than EMM 201. This increase in performance is because many of
the ABSs are set to CABSs providing a good coverage and capacity for the whole period. In
EMM 201, the ABSs are mostly in sleep mode when the offered traffic is low. This results in
more blocking until additional base stations are turned ON to meet the increase in traffic
demand.
Without topology management, the blocking probability is lower as expected since there is
better coverage throughout call duration.
0.12
No Topology Management
EMM 201
0.1 EMM 202
0.08
0.06
0.04
0.02
0
5 10 15 20 25 30
Constant Offered Traffic (Erlangs)
Figure 5-6: Blocking Probability against Constant Offered Traffic for different EMMs and no TM
Although the dropping probability in figure 5.7 for EMM 201 is quite high compared to that
of EMM 202 for 15 to around 25 Erlangs for the same reasons given for the blocking
probability, the overall dropping probability for the different schemes is about 5 to 6%
which is acceptable for a wireless system.
0.07
0.06 EMM 201
EMM 202
0.05
0.04
0.03
0.02
0.01
0
5 10 15 20 25 30
Figure 5-7: Dropping probability against Constant Offered Traffic for different EMMs and no TM
For dynamic offered traffic shown in figure 5.8, the blocking probability is fairly good with
EMM 201 scheme again giving the least performance because of the switching used in that
model. The EMM 202 scheme gives a better performance with 5% blocking which is close to
the no topology management scheme with about 4.5% blocking.
0.07
0.06 EMM 201
EMM 202
0.05
Block Probability
0.04
0.03
0.02
0.01
0
5 10 15 20 25 30
Dynamic Offered Traffic (Erlangs)
Figure 5-8: Blocking Probability against Dynamic Offered Traffic for different EMMs and no TM
Again, the dropping probability as seen in figure 5.9 is lower than that of blocking (figure
5.8) for both schemes including that of no topology management. EMM 202 also gives a
better performance in this case.
0.04
0.035 No Topology Management

EMM 201
EMM 202
0.03
0.025
0.02
0.015
0.01
0.005
0
5 10 15 20 25 30
Figure 5-9: Dropping Probability against Dynamic Offered Traffic for different EMMs and no TM
In figure 5.10, the comparison of both schemes in terms of saved energy shows that more
energy is saved using EMM 201 for high and low level of offered traffic. The decreasing
difference in the saved energy is due to increase in the number of users. Hence, there is a
corresponding increase in the number of active ABSs and energy consumed. No energy can
be saved when topology management is not in use because all the ABSs will remain ON
through the call period.
70%
60%
EMM 201
EMM 202
50%
40%
30%
20%
10%
0%
5 10 15 20 25 30
Figure 5-10: Percentage Energy Saved against Constant Offered Traffic for different EMMs and no TM
For dynamic offered traffic shown in figure 5.11, more energy is saved when either of the
two schemes is in use due to the varying traffic.
Since the CABSs used in EMM 202 scheme remains ON throughout the day, energy will be
wasted when it is not in use, while EMM 201 will be more energy efficient because ABSs
which are not in use or have low usage are turned to sleep mode.
EMM 201 can be argued to be a better scheme for the phase 2 because it gives a fairly good
balance between the blocking and dropping probabilities for both constant and dynamic
offered traffic based on the analysed performance.
Although in situations where the average blocking or dropping is desired to be as low as

possible, EMM 202 will be preferred, despite the lower performance in terms of energy
saved.
70%
EMM 201
60%
EMM 202
50%
40%
30%
20%
10%
0%
5 10 15 20 25 30
Figure 5-11: Percentage Energy Saved against Dynamic Offered Traffic for different EMMs and no TM
5.4. TEST AND ANALYSIS IN PHASE 3
5.4.1. SYSTEM EVALUATION USING PERFORMANCE MEASURES
Figure 5.12 shows that when the acceptance threshold is increased, there is a corresponding
increase in the probability of blocking because some users will have lower SINR values due
to their distance from respective ABSs. In the case where the acceptance threshold is too
high, users with lower SINR are blocked. Such users may find it difficult gaining access into
the system except when other users who act as interference to it are out of the system, or
farther away.
0.16
Threshold = 1.8 dB
Threshold = 3.6 dB
0.14
Threshold = 5.4 dB
0.12
0.1
0.08
0.06
0.04
0.02
0
40 50 60 70 80 90 100 110 120 130 140 150
Dynamic Offered Traffic(Erlangs)
Figure 5-12: Blocking Probability against Dynamic Offered Traffic for different thresholds using EMM 301
The plot in figure 5.13 also shows that the dropping probability increases as the acceptance
SINR value increases.
0.14
Threshold = 1.8 dB
Threshold = 3.6 dB
0.12 Threshold = 5.4 dB
0.1
0.08
0.06
0.04
0.02
0
40 50 60 70 80 90 100 110 120 130 140 150
Figure 5-13: Dropping Probability against Dynamic Offered Traffic for different thresholds using EMM 301
Comparing the results of the block and drop probabilities, it is observed that there is a
greater increase in dropping than in blocking. This is because of rise in the interference on
the system due to more users and ABSs. By closely observing the simulation, it is seen that
most of the drop and blocks occur in the access network because of the use of omni-
directional antennas. On the backhaul link directional antennas has been implemented
thereby reducing the interference from transmitters and a higher performance.
More energy is saved when the threshold is lowest because there will be no need to turn on
more base stations to meet rising traffic demand which results in greater energy
consumption.
35%
Threshold = 1.8 dB
Threshold = 3.6 dB
32.5
Threshold = 5.4 dB
30%
27.5%
25%
22.5%
20%
17.5%
15%
12.5%
40 50 60 70 80 90 100 110 120 130 140 150
Figure 5-14: Percentage Energy Saved against Dynamic Offered Traffic for different thresholds using EMM 301
Based on the analysis above, it is reasonable to use 3.8dB as the acceptance threshold since
it gives a good balance between dropping and blocking probabilities and saved energy.
5.4.2. COMPARISON OF TOPOLOGY MANAGEMENT SCHEMES
Only few results were produced for this part because due to time limitation, full
optimization of the code was not achieved.
Nevertheless, results from comparing EMM 301 and 302 in figure 5.15 shows that more
energy saving is achievable with the EMM 302 scheme because most of the ABSs are set to
SABSs and the CABSs are evenly positioned to provide an acceptable QoS.
It is obvious that less energy saving is achieved in EMM 301 scheme due to the fact that
more ABSs were ON. However, it provides better performance in terms of blocking and
dropping probabilities.
70%
EMM 301
EMM 302
60% No TM
50%
40%
30%
20%
10%
0%
40 50 60 70 80 90 100 110 120 130 140 150
Figure 5-15: Percentage Energy Saved against Dynamic Offered Traffic for EMM 301 &302
5.5. CONCLUSION
The dropping probability is often always seen to be lower than that of blocking as expected
because users with SINR lower than the set system threshold are not allowed into the
system and only a few ongoing calls which fall below the threshold will be dropped. In cases
where the blocking was very low in phase 3, the dropping was observed to increase.
The scheme from [53] gives better results in phase 2 in terms of blocking and dropping
probability, but lower energy savings. In phase 3, it gives a far better performance when
compared to EMM 301.
Increasing the number of channels in the backhaul and access network gives a far better
performance. We have only used a limited number of channels in the work presented
because of the capacity benefits in directionality on the backhaul links.
PROJECT MANAGEMENT 51
6. PROJECT MANAGEMENT
This project has been successfully completed because of management done from the start
to finish. Some important aspects on how the project was managed are discussed. A Gantt
chart showing how this project was organized till the end date it attached in the appendix.
Two supervisors were also assigned to monitor the project
David Grace was the primary supervisor, assisted by Paul Mitchell. They gave relevant
feedbacks for the first and second stages of this project and were prompt in responding to
issues that came up during the period of executing this project and writing the report.
Weekly meetings were held with our primary supervisor on a weekly basis throughout the
project.
6.1. DELIVERABLES
The deliverables written by Aisha Maukembayeva and updated from the second stage
report to help the team know well we have progressed, and what was accomplished at
different times is shown below.
Table 4: Table of deliverables; directly produced but updated from the second stage report
No Deliverables Date
1 First stage report 10/02/2014
2 Presentation 17/02/2014
3 Phase one 03/07/2014
4 Second stage report 03/07/2014
5 Phase two 01/08/2014
6 Phase three 25/08/2014
7 Submit final report 10/09/2014
8 Project VIVA 19/09/2014
6.2. RISK MANAGEMENT/MITIGATION
I have updated the risk management table from the second stage report. It highlights some
of the risks factors which would have been highly detrimental to the successful completion
of the project if not mitigated. The impact score is rated on a scale of 1 to 5 representing
minimum (low) to maximum (high) impact on the overall success of the project.
Table 5: Risk management
SN Risk Factor Reason Mitigation / Impact

score
Avoidance
01 Conflict or issues with -Work has been -Adequate code version 5

code integration. done individually control.
and with different
variable names. -Coordination between
group members.
Discrepancy in
parameters. -Agreement of
parameters.
02 Incorrect output after -Error in calculations -Proper and accurate 4

performance check. or conflicting codes. calculations.
-debugging of code.
03 Team member does -Lack of -To discuss problems 4

not meet deadline. understanding, with supervisor and team
error in code and members
calculations.
Other team members
-Team member is help out.
indisposed.
04 Error in cell -Incorrect -Again, care must be 3

dimensions. parameters. taken during parameter
calculation.
05 Team member exits the -Due to failed exams -Depending on the time 2
project before or lack of funds. of exit, his role will be
completion. excluded from the
project or transferred to
another team member
06 Challenges occur while -When he is on -Report the issue to the 4

supervisor is absent. vacation, or on an secondary supervisor.
assignment.
-Consult other members
of the PhD research team
working this area.
07 Loss of document. -Unexpected shut -Save files regularly. 5

down of computer
or word document. -Proper maintenance of
computer. Also restart
-Laptop or memory when necessary.
stick stolen.
-Use other secure
storage locations like the
sky drive.
08 Faulty personal -Due to adapter -Handle PC with care. 2

computer (PC) or issues or broken pc.
software. -There are lots of PC
-Unlicensed or available in the
outdated software. department get on a new
one.
-Report to IT services for

assistance.
09 Team member falls ill -Due to heavy -Take time to exercise. 1

or has to be away for a workload.
while. -Avoid procrastination
-Emergency and carry out task as
planned.
-Inform project chair and

supervisor before leaving
6.3. QUALITY ASSURANCE
In order to meet an acceptable standard, the risk factors in this project were taken into
consideration and some of the mitigations used are briefly discussed below [54]:
• System project risk assessment: Prompt response was given to potential risk areas of this
project which were identified under risk management. For example while our supervisor
was on vacation during the period of our project I had some challenges which were dealt
with by a member of the PhD research team. Also at some point we were behind the actual
plan, and had to review our Gantt chart.
• System project oversight: We have designed the network system in this project to
simulate real communication environment as close as possible, by different propagation
models, traffic model, channel assignment and energy management schemes.
• Testing and system acceptance support: Proper testing has been carried out at each phase
of this project. This involves ensuring that the codes written at every phase are in order and
that the integration of the various parts is done accurately. Although we have been unable
to get a reasonable result with the reinforcement learning scheme due to the limited time,
other issues regarding debugging of the code was achieved.
• Troubled project turn around: When some team members were unavailable due to health
challenges and examinations, it was promptly reported to our supervisor and the project
chair. As the project chair, I also confirmed that our supervisor was aware and made
necessary adjustments to keep the project running. We also referred to our Gantt chart and
set deliverables, to guide the team where necessary on how to progress and meet up with
set tasks.
• Improvements to project management and quality assurance: We had regular meetings

with our supervisor on a weekly basis and the team worked closely in the same
environment, in order to monitor the progress of other members and resolve any challenge
encountered. Changes in meeting time were promptly responded to by sending out email to
members of the team.
FURTHER WORK 55
7. FURTHER WORK
This project has covered the essential aspects of a telecommunication network. Although a
lot of work has been done, there are many more techniques to be considered in order to
achieve maximum energy with an acceptable balance on other performance measures.
Some of these techniques suggested for further work are:
 Improvement on topology management schemes: ABSs could be designed with the

ability to switch between topology management schemes. This will enable it to
adapt to its changing environment and give better performance. Current
development in digital signal processing, networking, machine learning, computer
software and hardware will make this possible with minimum delay [7].
 Use of antenna muting: As suggested in chapter 3, antenna ports on ABS not in use
at low traffic period can be muted, while boosting the power in those ABSs with
high load. This might serve as an alternative to the CABS suggested in [53].
 Cooperative Communication and Power Control: In situations where the user is very
far from the base station, future work can consider allowing the MS to adjust its
transmit power to get an acceptable QoS or reduce its power when it is close to the
base station, thereby saving more energy [33].
 Use of macro cell system: As proposed in [37], the HBS can be designed to operate
like a macro cell when the surrounding ABSs are turned off due to low offered
traffic. This will provide an acceptable coverage, and further increase the QoS.
 Directionality on the ABS: In this project, we have only applied directionality on the
backhaul in order to increase capacity and reduce interference issues. Nevertheless,
with directionality on the access network, interference issues will be further
reduced with a significant increase in the QoS.
 Use reinforcement learning: This was proposed in [55] suggesting ways in which we
can take advantage of distributed artificial intelligence by making use of
reinforcement learning with reduced complexity.
 Non-uniform user distribution: This project only considers the situation where the
users are uniformly distributed, but in [56] it is suggested that simulating with non-
uniform user distribution gives better real life scenarios.
CONCLUSION 56
8. CONCLUSION
In this report, the key features of the 5G network and the associating challenges have
highlighted. Solutions provided by other researchers and some techniques currently in use
were discussed together with the proposed technique later implemented in this work.
Our system was patterned after the BuNGee project, but with a lower level of complexity.
Constant and dynamic offered traffic were considered in all phases except phase three with
just the dynamic traffic. In the first phase, FSPL was used while the second and third phases
WINNER II B1 model was implemented for loss models. Directionality was also implemented
on the backhaul links but the access network used omni-directional antennas with different
channel assignment schemes considered for both access and backhaul networks.
In order to achieve energy saving which is a major area of concern in future

telecommunication systems, sleep mode technique was deployed with different suggested
methods of turning off and on ABSs using specified energy management rules. We have
achieved energy savings of about 35 to 60% for dynamic offered traffic and about 15 to 60%
for constant offered traffic in cases of high and low loads to confirm some of the results
achieved in [47] and the predicted result in [33].
Also, other measures like dropping and blocking probability were used to evaluate the
system performance. For both constant and dynamic traffic, a reasonable percent of
blocking and dropping was achieved. Comparisons were also made using different SINR
levels and energy management schemes.
ACKNOWLEDGEMENT 57
ACKNOWLEDGEMENT
My sincere appreciation goes to my primary supervisor Dr. David Grace for guiding my team
and me, from the start to completion of the project. He provided the materials needed at
the initial stage of this work which was vital in positioning the project on course, and
promptly response was given to various challenges that came up during this period.
Secondly, my secondary supervisor Dr. Paul Mitchell who also provided relevant answers to
the questions I had at some point during the project. The feedback he gave at the first and
second of the project was also supportive.
I also want to thank my team mates Aisha Maukembayeva, Ao Lei and Ogheneovie Ajemuta
for their support throughout this period, not forgetting Yunbo Han and the Communications
Research Group here at The University of York.
Finally, my appreciation goes to my parents Mr. & Mrs. Friday Jakpa who gave me the
opportunity to study abroad, and to God for all the provisions.
REFERENCES 58
REFERENCES
[1] M. Goldhamer et al, “BuNGee Deliverables: D2.1 Baseline BuNGee Architecture,” p. 18,
09 November 2010.
[2] A. S. A. Latef and R. Hassan, “Spectrum Management System: A Study,” in International

Conference on Electrical Engineering and Informatics, Bandung, Indonesia, July, 2011.
[3] Huawei, Whitepaper on Spectrum, Shenzhen: Huawei Technologies Co., Ltd, 2013.
[4] K. Pretz, “Overcoming Spectrum Scarcity,” The IEEE news source, 13 August 2012.
[Online]. Available: http://theinstitute.ieee.org/technology-focus/technology-
topic/overcoming-spectrum-scarcity. [Accessed 26 August 2014].
[5] “The UK Spectrum Strategy,” Department for Culture Media & Sport, 2014.
[6] Federal Communications Commission, “Notice of Proposed Rule Making and Order,” ET
Docket no. 03-322, Washington, D.C., 2003.
[7] S. Haykin, “Cognitive Radio: Brain-Empowered Wireless Communications,” IEEE Journal

on selected areas in communications, vol. 23, pp. 201-220, Feb,2005.
[8] L. Yigal et al, “BuNGee Deliverable: D3.1 Baseline RRM & Joint Access/Self-Backhaul
Design,” 2011.
[9] J. Mitola and G. Q. Maguire, “Cognitive Radio: Making Software Radios More Personal,”
IEEE Personal Communications, vol. 6, no. 4, pp. 13-18, August 1999.
[10] M. Goldhamer et al, “BuNGee Deliverable: D1.3 Final BuNGee Architecture,” January,
2012.
[11] “BuNGee Deliverable: D1.2 Baseline BuNGee Architecture”.
[12] P. Sharma, “Evolution of Mobile Wireless Communication Networks - 1G to 5G as well

as Future Prospective of Next Generation Communication Network,” International
Journal of Computer Science and Mobile Computing, vol. 2, no. 8, pp. 47-53, 2013.
REFERENCES 59
[13] T. Jiang et al, “BuNGee Deliverable: D4.1.1 Interim Simulation,” 2011.
[14] Y. Han, “Energy Efficient Topology Management for Beyond Next Generation Mobile
Broadband Systems (Thesis),” Communication Research Group, University of York,
September, 2013.
[15] C. Han, T. Harrold, S. Armour, T. Anh and M. R. Nakhai, “Green Radio: Radio Techniques
to Enable Energy-Efficient Wireless Networks,” Communications Magazine, IEEE, May
2011.
[16] O. Tamm, C. Hermsmeyer and A. M. Rush, “Eco-Sustainable System and Network

Architectures for Future Transport Networks,” Bell Labs Technical Journal, vol. 14, no. 4,
pp. 311-328, 2010.
[17] K. Hinton et al, “Power Consumption and Energy Efficiency in the Internet,” IEEE
Network, vol. 25, no. 2, pp. 6-12, March 2011.
[18] GESI, “Smart 2020: Enabling the low Carbon Economy in the Information Age,” The
Climate Group, 2008.
[19] M. Walker, “Surfing the Green Wave in Telecom,” Ovum, 9 May 2008.
[20] Mobile VCE, “Green Radio,” 2013. [Online]. Available:

http://www.mobilevce.com/green-radio. [Accessed 8 August 2014].
[21] M. Deruyck et al, “Reducing the Power Consumption in Wireless Access Networks:
Overview and Recommendations,” Progress In Electromagnetics Research, vol. 132, pp.
255-274, 2012.
[22] K. Kawatomo et al, “Electricity Used by Office Equipment and Network Equipment in
the US,” in ACEEE Summer Study Conf. Energy Efficiency in Buildings (NLBL-45917),
2000.
[23] J. Baaliga et al, “Energy Consumption in Optical IP Networks,” Lightwave Technology

Journal, vol. 27, no. 13, pp. 2391-2403, 2009.
[24] D. C. Kilper et al, “Power Trends in Communications Networks,” IEEE Journal of Selected
REFERENCES 60
Topics in Quantum Electronics, vol. 17, no. 2, pp. 275-284, March 2011.
[25] X. Xu, K. Gledi and M. Rudolf, “Energy Efficient Power Management for 4G
Heterogeneous Cellular Networks,” in 1st International Workshop on Green Optimised
Wireless Networks (GROWN), IEEE, 2013.
[26] A. Damajanovic et al, “A Survey on 3GPP Heterogeneous Networks,” IEEE Wireless

Communications, vol. 18, no. 3, pp. 10-21, June 2011.
[27] Q. Li et al, “MIMO Tecchniques in WiMAX and LTE: a feature overview,”

Communications Magazine, IEEE, vol. 48, no. 5, pp. 86-92, May 2010.
[28] C. Han and S. Amour, “Energy Efficient Radio Resource Management Strategies for
Green Radio,” IET Communications, July 2011.
[29] W. Cheng, H. Zhang, L. Zhao and Y. Li, “Energy Efficient Spectrum Allocation for Green
Radio in Two-Tier Cellular Networks,” in Global Telecommunications Conference
(GLOBECOM), IEEE, Miami, USA, December, 2010.
[30] P. Skillermark and F. Pal, “Enhancing Energy in LTE with Antenna Muting,” in Vehicular
Technology Conference (VTC Spring), IEEE, Yokohama, 2012.
[31] P. Frenger, H. Koorapaty and J. Guey, “Evaluation of COntrol Channel Performance with
Adaptive Radio Unit Activation in LTE,” in IEEE Vehicular Technology Conference (VTC
Spring), Yokohama, Japan, 2012.
[32] I. Ashraf, L. Ho and H. Claussen, “Improving Energy Efficiency of Femtocell Base Stations
Via User Activity Detection,” in IEEE Wireless Communications and Networking
Conference (WCNC), Sydney, Australia, April, 2010.
[33] F. Han, Z. Safar, S. W. Lin, Y. Chen and R. K. J. Liu, “Energy-Efficient Cellular Network
Operation via Base Station Cooperation,” in IEEE International Conference on
Communications (ICC), June, 2012.
[34] S. Landstorm et al, “Heterogenous Networks - Increasing Cellular Capacity,” [Online].

Available:
http://www.ericsson.com/res/thecompany/docs/publications/ericsson_review/2011/h
REFERENCES 61
eterogeneous_networks.pdf. [Accessed 06 08 2014].
[35] C. Khirallah and J. Thompson, “Energy and Cost Impact of Relay and Femtocell
Deployments in LTE-Advanced,” IET Communications, vol. 5, no. 18, pp. 2617-2628, 16
December 2011.
[36] W. Guo and T. O'Farrell, “Relay Deployment in Cellular Networks: Planning and
Optimization,” IEEE Journal on Selected Areas in Communications (JSAC), vol. 31, no. 8,
pp. 1597-1606, August 2013.
[37] S. Rehan and D. Grace, “Macro-Cell Overlaid Green Topology Management for Beyond
Next Generation Mobile Broadband Systems,” in Conference Proceedings at The
University of York, Future Network & Mobile Summit, IIMC, 2013.
[38] W. Guo, S. Wang, T. O'Farrell and S. Fletcher, “Energy Consumption of 4G Cellular

Networks: A London Case Study,” in IEEE Vehicular Technology Conference (VTC Spring),
Dresden, 2013.
[39] J. Andrusenko, J. Burbank and J. Ward, “Modeling and Simulation for RF Propagation,”
[Online]. Available:
http://cms.comsoc.org/SiteGen/Uploads/Public/Docs_Globecom_2009/3-
propagationM_S.pdf. [Accessed 02 09 2014].
[40] RAYmaps, “WINNER-II Path Loss Model,” 06 01 2012. [Online]. Available:

http://www.raymaps.com/index.php/winner-ii-path-loss-model/. [Accessed 03 09
2014].
[41] Y. Chen, S. Zhang, S. Xu and G. Li, “Fundamental Tradeoffs on Green Wireless

Networks,” IEEE Communications Magazine, vol. 49, no. 6, pp. 30-37, 2011.
[42] Jemdoc, “Signal to Interefence plus Noise Ratio(SINR) in Wireless Communications,” 13

01 2014. [Online]. Available:
https://inst.eecs.berkeley.edu/~ee127a/book/login/exa_gp_power_wless.html.
[Accessed 01 09 2014].
[43] P. Mitchell, “Module ELE00012H: Mobile Communications Systems,” [Online].
REFERENCES 62
Available:
http://www.elec.york.ac.uk/internal_web/meng/yr3/modules/Mobile_Comms_System
s/MCS2013_StudentNotes24-43.pdf. [Accessed 01 09 2014].
[44] G. Wagoner, “Traffic Engineering: Blocking Probability,” [Online]. Available:

http://garywagoner.tripod.com/traffic/BLOCKING.HTML. [Accessed 01 09 2014].
[45] N. S. Tarkaa, J. M. Morn and C. I. Ani, “Drop Call Probability Factors in Cellular
Networks,” International Journal of Sceintific & Engineering Research, vol. 2, no. 10, pp.
1-5, October, 2011.
[46] N. Shi, J. Zheng, B. Xin and Y. Ge, “Energy Model for GSM Base Station,” China Mobile
Communication, vol. 32, no. 2, 2011.
[47] Y. Han, D. Grace and P. Mitchell, “Energy Efficient Topology Management for Beyond
Next Generation Systems,” in IEEE International Symposium on Wireless
Communication Systems (ISWCS), Paris, August, 2012.
[48] S. J. Thompson, “Green Radio Research Overview,” Department of Electronics, The

University of York, pp. 4-5, May, 2014.
[49] C. S. Bontu and E. Illidge, “DRX Mechanism for Power Saving in LTE,” IEEE
Communications Magazine, vol. 47, no. 6, pp. 48-55, June, 2009.
[50] D. Grace, “Using Cognition to Deliver 'Green' Communications,” GREENTIC Worshop,

Casablanca, Morocco, November, 2011.
[51] A. Fisusi, D. Grace and P. Mitchell, “Energy Efficient Cluster-Based Resource Allocation
and Topology Management for Beyond Next Generation Mobile Broadband Networks,”
in IEEE International Conference on Communications Workshops (ICC), Budapest, 2013.
[52] T. Srinidhi, G. Sridhar and V. Sridhar, “Topology Management in Ad Hoc Mobile

Wireless Networks,” in Applied Research Group Satyam Computer Services Ltd.,
Bangalore 560 025, India, 2006.
[53] L. Zhou, “Cognitive Radio for 4G Energy Efficient Communications,” for MSc Project at
The University of York, August, 2012.
REFERENCES 63
[54] Deloitte, “Project Quality Assurance,” 2014. [Online]. Available:

http://www.deloitte.com/view/en_GR/gr/services/enterprise-risk-services/project-
quality-assurance/. [Accessed 20 June 2014].
[55] D. Grace, J. Chen, T. Jiang and P. D. Mitchell, “Using Cognitive Radio to Deliver 'Green'
Communications,” in Proceedings of the 4th International Conference on COWNCOM,
2009.
[56] R. Ganesh and J. Kuriacose, “Effect of Non-Uniform Traffic Distribution on Performance

of a Cellular CDMA System,” in IEEE 6th INternational Conference on Universal Personal
Communications Record, 1997.
APPENDIX A 64
APPENDIX A
Time line for this project
APPENDIX B 65
APPENDIX B
% 'CR 5G Network' - Phase Two-System Intergate V3.1
% Entire 5G CR Network, Ogheneovie Ajemuta, Aisha Muakembayeva, Oyomikon Ezekiel Jakpa, Ao Lei
%% Parameter initialisation
BW = 10e6; % FDD for only uplink
Bolz_c = 1.38e-23; % Bolzman Conatant
T_kelv = 295; % Temperature in Kelvin
N = BW*Bolz_c*T_kelv;
% transmit power, dBm -> dBw
pt_BS_dB = 38 - 30;
pt_MS_dB = -7;
E_ABS = 5;
E_wakeup = 50;
E_spare = 0.25;
% Number of communication stations
num_MS = 1000;
num_HSS = 76;
num_ABS = 76;
num_HBS = 5;
num_CH_acc = 6; % Number of channel for access network
num_CH_bh = 4; % Number of channel for backhaul network
SNIR_thres = 10^(1.8/10); % SNIR threshold
Block_num = 0; % Number of call was blocked
Drop_num = 0; % Number of call was dropped
% Generate Propagation and Traffic Model
[GC_MS_ABS_dB,GC_MS_HBS_dB,GC_HSS_HBS_dB,GC_HSS_ABS_dB]=PM_P3(num_MS);
% [t_Arr,t_Dep]=Traffic_C_sec(num_MS);
[t_Arr,t_Dep]=Traffic_D_sec(num_MS,8);
% Sort Event
% (figure out if two call at same time) dosen't matter, don't need this step
[t_Arr_sort,t_Arr_seq] = sort(reshape(t_Arr,1,length(t_Arr(:,1))*length(t_Arr(1,:))));
t_Arr_seq = t_Arr_seq((length(find(t_Arr_sort==0))+1):end); % it must BEFORE next step!
t_Arr_seq = t_Arr_seq - (ceil(t_Arr_seq/num_MS)-1)*num_MS; % Which MS at Which Time
[t_Dep_sort,t_Dep_seq] = sort(reshape(t_Dep,1,length(t_Dep(:,1))*length(t_Dep(1,:))));
t_Dep_seq = t_Dep_seq((length(find(t_Dep_sort==0))+1):end); % it must BEFORE next step!
t_Dep_seq = t_Dep_seq - (ceil(t_Dep_seq/num_MS)-1)*num_MS; % Which MS at Which Time
% t_seq -> t_xxx_seq (find back)
t_sort(1:length(t_Arr_seq)) = t_Arr_sort((length(find(t_Arr_sort==0))+1):end);
t_sort((length(t_Arr_seq)+1):(2*length(t_Arr_seq))) =
t_Dep_sort((length(find(t_Dep_sort==0))+1):end);
[t_sort,t_seq] = sort(t_sort);
Drop_num_acc = zeros(1,length(t_sort));
Drop_num_bh = zeros(1,length(t_sort));
Block_num_acc = zeros(1,length(t_sort));
Block_num_bh = zeros(1,length(t_sort));
seq_mark = 1;
alloff = 0;
num_on = 0;
% Block_num_acc = zeros(1,length(t_sort));
% Transfer from dB to Linear
pt_BS = 10^(pt_BS_dB/10);
pt_MS = 10^(pt_MS_dB/10);
GC_MS_ABS = 10.^(GC_MS_ABS_dB./10);
GC_HSS_HBS = 10.^(GC_HSS_HBS_dB./10);
% Matrices
ABS_status = ones(1,num_ABS); % ABS on/off status
ABS_OnOff = ones(num_ABS,length(t_sort));
SNIR_acc = zeros(num_ABS,num_CH_acc);
SNIR_bh = zeros(num_HBS,num_CH_bh);
CH_MS_acc = zeros(num_MS,num_CH_acc); % Initialise access network channel occupy matrix
CH_ABS_acc = zeros(num_ABS,num_CH_acc); % Channel occupied by ABSs
CH_HSS_bh = zeros(num_HSS,num_CH_bh); % Initialise backhaul channel occupy matrix
CH_MS_bh = zeros(num_MS,num_CH_bh); % Which MS use which channel
MS_ABS_conn = zeros(num_HSS,num_MS); % Which MS connects to which ABS
ABS_nuconn = 0;
% ABS Energy/Traffic Level
E_Lv = [1,10,67,76,5,6,23,35,42,54,71,72,2,4,7,9,11:14,19:22,24,26,28,30,32,34,...
36,37,40,41,43,45,47,49,51,53,55:58,63:66,68,70,73,75]; % From E level high to low, 56
ABSs
E_adj(1,:) = [2,9,63,66,4,7,19,22,43,53,64,65,1,3,6,8,1,5,6,10,15,16,17,18,23,25,27,29,31,...
33,27,31,38,39,42,44,46,48,50,52,42,46,50,54,59,60,61,62,67,69,72,74]; % Adjacent ABS
table row 1
APPENDIX B 66
E_adj(2,:) = [11,14,68,75,12,13,24,34,55,58,70,73,3,5,8,10,15,16,17,18,23,27,31,35,25,27,...
29,31,33,35,38,39,46,50,44,46,48,50,52,54,59,60,61,62,67,71,72,76,69,71,74,76]; % Adjacent
ABS table row 2
% Parameter initialisation end
%% MAIN FRAME
tic
for i=1:length(t_sort) % This loop is used when whole system is integrated
%% ACCESS NETWORK
% % By Ogheneovie Ajemuta
% % THIS PART IS FOR ACCESS NETWORK AND INPUT FOR BACKHAUL NETWORK
MS_ABS_past = MS_ABS_conn; % ?????
% % If call Departure, release channel
if t_seq(i)>(length(t_sort)/2) % >length(t_sort)/2 -> Dep
% Which MS's call departures
call_dep_no = t_Dep_seq(t_seq(i)-(length(t_sort)/2));
CH_dep = ceil(find(CH_MS_acc(call_dep_no,:)==1)/length(call_dep_no)); % Which channel is
released
CH_MS_acc(call_dep_no,:) = 0; % Release Access network channel
% MS_ABS_conn(:,call_dep_no(n))==1 is the MS correspond to which ABS
CH_ABS_acc(MS_ABS_conn(:,call_dep_no)==1,CH_dep) = 0; % it must BEFORE next step!
MS_ABS_conn(MS_ABS_conn(:,call_dep_no)==1,call_dep_no) = 0;
else
% DROP OR NOT
if i~=1 && isempty(find(CH_ABS_acc==1,1))~=1
% With condition 2: 0.081285 s, without: 0.101881 s
CH_use_acc = ceil(find(CH_ABS_acc==1)/num_ABS);
ABS_use_acc = find(CH_ABS_acc==1) - num_ABS*(CH_use_acc - 1);
for j=1:length(ABS_use_acc)
% Current SNIR calculations (Real SNIR)
MS_using = find(MS_ABS_conn(ABS_use_acc(j),:)~=0);
MS_use = MS_using(ismember(MS_using,find(CH_MS_acc(:,CH_use_acc(j))~=0)));
GC_int_drop = GC_MS_ABS(ABS_use_acc(j),:)*CH_MS_acc(:,CH_use_acc(j))...
- GC_MS_ABS(ABS_use_acc(j),MS_use)*CH_MS_acc(MS_use,CH_use_acc(j));
SNIR_drop = pt_MS*GC_MS_ABS(ABS_use_acc(j),MS_use)/(N + pt_MS*GC_int_drop);
if SNIR_drop<SNIR_thres
CH_ABS_acc(ABS_use_acc(j),CH_use_acc(j)) = 0;
CH_MS_acc(MS_use,CH_use_acc(j)) = 0;
MS_ABS_conn(ABS_use_acc(j),MS_use) = 0;
Drop_num_acc(i) = 1;
% Release Backhaul channels
CH_HSS_bh(ABS_use_acc(j),CH_MS_bh(MS_use,:)==1) = 0;
CH_MS_bh(MS_use,:) = 0;
end
end
end
% Drop or not ends
MS_call = t_Arr_seq(t_seq(i));
ABS_valid = find(bsxfun(@and,(sum(CH_ABS_acc')<=4),ABS_status)==1);
SNIR_acc = zeros(length(ABS_valid),num_CH_acc);
for j=1:length(ABS_valid) % ABS number
for k=1:num_CH_acc % Channel number
% Calculate Combined interference gain coefficient (Assumed SNIR)
GC_int_acc = GC_MS_ABS(ABS_valid(j),:)*CH_MS_acc(:,k) -
GC_MS_ABS(ABS_valid(j),MS_call)*CH_MS_acc(MS_call,k);
SNIR_acc(j,k) = pt_MS*GC_MS_ABS(ABS_valid(j),MS_call)/(N + pt_MS*GC_int_acc); %
Calculate SNIR
end
end
SNIR_acc = SNIR_acc.*(1-CH_ABS_acc(ABS_valid,:));
SNIR_max = max(max(SNIR_acc));%.*(1-CH_ABS_acc(ABS_valid,:)))); % Max SNIR, elimiate a
channel in same ABS using multi channels.
if SNIR_max>=SNIR_thres % Compare with threshold
SNIR_max_no = find(SNIR_acc==SNIR_max); % ABS*Channel with max SNIR
rand_conn_no = SNIR_max_no(randi(length(SNIR_max_no))); % Pick channel randomly
CH_conn_no = ceil(rand_conn_no/length(ABS_valid)); % Calculate Channel number
ABS_conn_no = ABS_valid(rand_conn_no - length(ABS_valid)*(CH_conn_no-1)); % Calculate
ABS number
CH_MS_acc(MS_call,CH_conn_no) = 1; % Update channel occupied matrix
CH_ABS_acc(ABS_conn_no,CH_conn_no) = 1;
MS_ABS_conn(ABS_conn_no,MS_call) = 1; % Update which MS connect with which ABS
else
Block_num_acc(i) = 1;
end
end
% ACCESS NETWORK END
APPENDIX B 67
%% ENERGY SAVING MODE

% % By Oyomikon Ezekiel Jakpa
% How many times MS connects
ABS_nuconn = sum(MS_ABS_conn') + ABS_nuconn;
% ENERGY SAVING MODE END
%% BACKHAUL NETWORK
% % By Aisha Muakembayeva
% THIS PART IS FOR BACKHAUL NETWORK
if t_seq(i)>(length(t_sort)/2) % >length(t_sort)/2 -> Dep
% Which MS's call departures
call_dep_no = t_Dep_seq(t_seq(i)-(length(t_sort)/2));
CH_HSS_bh(MS_ABS_past(:,call_dep_no)==1,CH_MS_bh(call_dep_no,:)==1) = 0;
CH_MS_bh(call_dep_no,:) = 0;
else
% DROP OR NOT
if i~=1 && isempty(find(CH_HSS_bh==1,1))~=1
CH_use_bh = ceil(find(CH_HSS_bh==1)/num_ABS);
HSS_use_bh = find(CH_HSS_bh==1) - num_ABS*(CH_use_bh - 1);
for j=1:length(HSS_use_bh)
for k=1:num_HBS
% Current SNIR calculations
MS_using = find(MS_ABS_conn(HSS_use_bh(j),:)~=0);
MS_use_bh = MS_using(ismember(MS_using,find(CH_MS_bh(:,CH_use_bh(j))~=0)));
GC_int_drop = GC_HSS_HBS(HSS_use_bh(j),:,k)*CH_HSS_bh(:,CH_use_bh(j))...
-
GC_HSS_HBS(HSS_use_bh(j),HSS_use_bh(j),k)*CH_HSS_bh(HSS_use_bh(j),CH_use_bh(j));
SNIR_drop(k) = pt_BS*GC_HSS_HBS(HSS_use_bh(j),HSS_use_bh(j),k)/(N +
pt_BS*GC_int_drop);
end
SNIR_drop = max(SNIR_drop); % Pick a max SNIR between HBS#1 and #2
if SNIR_drop<SNIR_thres
CH_HSS_bh(HSS_use_bh(j),CH_use_bh(j)) = 0;
CH_MS_bh(MS_use_bh,CH_use_bh(j)) = 0;
% Release Acc Channels when Bh drop happens
CH_ABS_acc(HSS_use_bh(j),CH_MS_acc(MS_use_bh,:)==1) = 0;
CH_MS_acc(MS_use_bh,:) = 0;
MS_ABS_conn(HSS_use_bh(j),MS_use_bh) = 0;
Drop_num_bh(i) = 1;
end
end
end
% Drop or not ends
% Channel Assignment
if Block_num_acc(i)~=1 % Acc blocked, then no need to assign Bh channel
for j=1:num_CH_bh % Channel number
for k=1:num_HBS
% Calculate Combined interference gain coefficient
GC_Int_bh = GC_HSS_HBS(ABS_conn_no,:,k)*CH_HSS_bh(:,j) ...
- GC_HSS_HBS(ABS_conn_no,ABS_conn_no,k)*CH_HSS_bh(ABS_conn_no,j);
SNIR_bh(k,j) =
pt_BS*GC_HSS_HBS(ABS_conn_no,ABS_conn_no,k)/(N+pt_BS*GC_Int_bh); % Calculate SNIR
end
end
% Select which HBS to Connect
SNIR_bh = bsxfun(@times,SNIR_bh,(1-CH_HSS_bh(ABS_conn_no,:))); % Valid SNIR
SNIRmax = find(sum(SNIR_bh')==max(sum(SNIR_bh'))); % SNIR max or which HBS connects
SNIR_bh = SNIR_bh(SNIRmax(randi(length(SNIRmax))),:); % Valid SNIR vector
SNIR_max_bh = max(SNIR_bh);%.*(1-CH_HSS_bh(ABS_conn_no,:))); % Max SNIR
if SNIR_max_bh>=SNIR_thres % Compare with threshold
SNIR_max_bh_no = find(SNIR_bh==SNIR_max_bh); % ABS*Channel with max SNIR
rand_conn_no = SNIR_max_bh_no(randi(length(SNIR_max_bh_no))); % Pick channel
randomly
CH_conn_no = rand_conn_no; % Calculate Channel number
CH_HSS_bh(ABS_conn_no,CH_conn_no) = 1; % Update channel occupied matrix
CH_MS_bh(MS_call,CH_conn_no) = 1;
else
Block_num_bh(i) = 1;
% Backhaul blocked, release corresponding Acc channel as well
CH_MS_acc(MS_call,CH_conn_no) = 0; % Release channel occupied matrix
CH_ABS_acc(ABS_conn_no,CH_conn_no) = 0; % Release ABS-Channel matrix
MS_ABS_conn(ABS_conn_no,MS_call) = 0; % Release MS-ABS matrix
end
end
end
% BACKHAUL NETWORK END
%% ENERGY SAVE MODE
% % By Oyomikon Ezekiel Jakpa
APPENDIX B 68
CH_cap{i} = sum(CH_ABS_acc');
t_int = t_sort(i) - t_sort(seq_mark); % t'-t
if t_int>=10
Cap_ABS = zeros(1,num_ABS);
for j=1:(i-seq_mark+1)
% Capacity Calculation
Cap_ABS = Cap_ABS + CH_cap{j};
end
% Max Capacity value in BH channel number but not ACC channel number
Cap_ABS = Cap_ABS/(i-seq_mark+1); % Average Capacity, at t'-t
Blocknum = sum(Block_num_acc(seq_mark:i)) + sum(Block_num_bh(seq_mark:i)); % Block is not
considered
% SWITCH OFF
for j=1:length(E_Lv)
% Cap_t = CH_cap{seq_mark}; % CH_cap{seq_mark} is the capacity at t
% sum((CH_cap{seq_mark}(E_adj(:,j)))<(num_CH_bh*0.5))==2 capacity of all adj ABSs at t
<50%
% mean(Cap_ABS(E_adj(:,j))<(num_CH_bh*0.5)) average capacity at t-t' <50%
% sum(ABS_status(E_adj(:,j))) = all adjacents working
if sum(ABS_status(E_adj(:,j)))==2 && CH_cap{seq_mark}(E_Lv(j))<(num_CH_bh*0.25) &&
Cap_ABS(E_Lv(j))<(num_CH_bh*0.25) && ...
sum((CH_cap{seq_mark}(E_adj(:,j)))<(num_CH_bh*0.25))==2 &&
mean(Cap_ABS(E_adj(:,j))<(num_CH_bh*0.25)) && Blocknum==0
ABS_status(E_Lv(j)) = 0;
ABS_OnOff(E_Lv(j),i:end) = 0;
elseif (sum((CH_cap{seq_mark}(E_adj(:,j)))>=(num_CH_bh*0.25))==2 &&
mean(Cap_ABS(E_adj(:,j))>=(num_CH_bh*0.25))) || Blocknum>0
ABS_status(E_Lv(j)) = 1;
ABS_OnOff(E_Lv(j),i:end) = 1;
end
end
seq_mark = i;
end
% ENERGY SAVE MODE END
end
E_cost = Energy_Calculator(ABS_OnOff,num_ABS,t_sort);
figure(2)
bar(ABS_nuconn);
Blockprob = (sum(Block_num_acc) + sum(Block_num_bh))/(length(t_sort)/2)

Blockprob_acc = sum(Block_num_acc)/(length(t_sort)/2)
Blockprob_bh = sum(Block_num_bh)/(length(t_sort)/2)
Dropprob = (sum(Drop_num_acc) + sum(Drop_num_bh))/(length(t_sort)/2)
Dropprob_acc = sum(Drop_num_acc)/(length(t_sort)/2)
Dropprob_bh = sum(Drop_num_bh)/(length(t_sort)/2)
E_cost_percent = E_cost/((E_ABS*24*60*60 + E_wakeup)*num_ABS)
timecost1 = toc

Cognitive Radio For 5G Small - Final Report Done

Uploaded by

Copyright:

Available Formats

Cognitive Radio For 5G Small - Final Report Done

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Cognitive Radio For 5G Small - Final Report Done

Uploaded by

Copyright:

Available Formats

Department of Electronics

COGNITIVE RADIO FOR 5G SMALL-CELL

MSc. in Communications Engineering

Oyomikon Ezekiel Jakpa

Primary Supervisor: Dr. David Grace

Second Supervisor: Dr. Paul Mitchell

10th September, 2014

CONTENTS ................................................................................................................................. iii

LIST OF ACRONYMS ................................................................................................................... vi

LIST OF FIGURES ....................................................................................................................... vii

LIST OF TABLES .......................................................................................................................... ix

1.1. Overview ..................................................................................................................... 1

1.2. REPORT OUTLINE......................................................................................................... 3

2. BACKGROUND STUDIES ..................................................................................................... 4

2.1. INTRODUCTION ........................................................................................................... 4

2.2. COGNITIVE RADIO ....................................................................................................... 4

2.3. BEYOND NEXT GENERATION (BuNGee) PROJECT ....................................................... 5

2.4. CONCLUSION ............................................................................................................... 8

3. LITERATURE REVIEW .......................................................................................................... 9

3.1. INTRODUCTION ........................................................................................................... 9

3.2. WHY ENERGY MANAGEMENT? ................................................................................... 9

3.3. GREEN COMMUNICATIONS ...................................................................................... 10

3.4. APPROACHES FOR ENERGY EFFICIENCY .................................................................... 12

3.5. CONCLUSION ............................................................................................................. 14

4. SYSTEM DESIGN AND IMPLEMENTATION ........................................................................ 15

4.1. INTRODUCTION ......................................................................................................... 15

4.2. SYSTEM STRUCTURE .................................................................................................. 15

4.2.1. PHASE 1 .............................................................................................................. 16

4.2.2. PHASE 2 .............................................................................................................. 17

4.2.3. PHASE 3 .............................................................................................................. 17

4.2.4. TRAFFIC AND PROPAGATION MODEL ................................................................ 19

4.2.5. ACCESS NETWORK ............................................................................................. 20

4.2.6. BACKHAUL NETWORK ........................................................................................ 20

4.3. SYSTEM LEVEL ........................................................................................................... 21

4.4. SYSTEM PERFORMANCE MEASURES ......................................................................... 22

4.5. SYSTEM ENERGY MANAGEMENT .............................................................................. 26

4.5.1. ENERGY MANAGEMENT .................................................................................... 26

4.5.2. ENERGY MANAGEMENT RULES ......................................................................... 28

4.5.3. ENERGY EQUATIONS .......................................................................................... 29

4.6. ENERGY MANAGEMENT MODEL (EMM) 101 ............................................................ 32

4.7. ENERGY MANAGEMENT MODEL (EMM) 201 – 202 .................................................. 32

4.7.1. EMM 201 ............................................................................................................ 32

4.7.2. EMM 202 ............................................................................................................ 33

4.8. ENERGY MANAGEMENT MODEL (EMM) 301 – 302 .................................................. 34

4.8.1. EMM 301 ............................................................................................................ 34

4.8.2. EMM 302 ............................................................................................................ 35

4.9. CONCLUSION ............................................................................................................. 38

5. SYSTEM TEST AND ANALYSIS ........................................................................................... 39

5.1. INTRODUCTION ......................................................................................................... 39

5.2. SYSTEM CHECK .......................................................................................................... 39

5.3. TEST AND ANALYSIS IN PHASE 2 ............................................................................... 40

5.3.1. SYSTEM EVALUATION USING PERFORMANCE MEASURES ................................ 40

5.3.2. COMPARISON OF TOPOLOGY MANAGEMENT SCHEMES .................................. 43

5.4. TEST AND ANALYSIS IN PHASE 3 ............................................................................... 47

5.4.1. SYSTEM EVALUATION USING PERFORMANCE MEASURES ................................ 47

5.4.2. COMPARISON OF TOPOLOGY MANAGEMENT SCHEMES .................................. 49

5.5. CONCLUSION ............................................................................................................. 50

6. PROJECT MANAGEMENT ................................................................................................. 51