Power Control and Energy Efficiency Strategies for D2D ... · Em outras palavras, a tecnologia D2D pode ser utilizada para aumentar ainda mais a eficiência de espectro e a eficiência - [PDF Document] (2024)

FEDERAL UNIVERSITY OF CEARÁ

DEPARTMENT OF TELEINFORMATICS ENGINEERING

POSTGRADUATE PROGRAM IN TELEINFORMATICS ENGINEERING

Power Control and Energy Efficiency Strategies for

D2D Communications Underlying Cellular Networks

Master of Science Thesis

Author

Yuri Victor Lima de Melo

Advisor

Prof. Dr. Tarcisio Ferreira Maciel

Co-Advisor

Prof. Dr. Emanuel Bezerra Rodrigues

FORTALEZA – CEARÁ

JULY 2015

UNIVERSIDADE FEDERAL DO CEARÁ

DEPARTAMENTO DE ENGENHARIA DE TELEINFORMÁTICA

PROGRAMA DE PÓS-GRADUAÇÃO EM ENGENHARIA DE TELEINFORMÁTICA

Controle de potência e estratégias de eficiência

energética para comunicações D2D subjacentes redes

celulares

Autor

Yuri Victor Lima de Melo

Orientador

Prof. Dr. Tarcisio Ferreira Maciel

Co-orientador

Prof. Dr. Emanuel Bezerra Rodrigues

Dissertação apresentada à Coordenação do

Programa de Pós-graduação em Engenharia

de Teleinformática da Universidade Federal

do Ceará como parte dos requisitos para

obtenção do grau de Mestre em Engenharia

de Teleinformática. Área de concentração:

Sinais e sistemas.

FORTALEZA – CEARÁ

JULHO 2015

Dados Internacionais de Catalogação na Publicação Universidade Federal do Ceará

Biblioteca de Pós-Graduação em Engenharia - BPGE

M78p Melo, Yuri Victor Lima de.

Power control and energy efficiency strategies for D2D communications underlying cellular networks / Yuri Victor Lima de Melo. – 2015.

72 f. : il. color. , enc. ; 30 cm. Dissertação (mestrado) – Universidade Federal do Ceará, Centro de Tecnologia, Departamento de

Engenharia de Teleinformática, Programa de Pós-Graduação em Engenharia de Teleinformática, Fortaleza, 2015.

Área de concentração: Sinais e Sistemas. Orientação: Prof. Dr. Tarcísio Ferreira Maciel. Orientação: Prof. Dr. Emanuel Bezerra Rodrigues. 1. Teleinformática. 2. Controle de potência. 3. Interferência - Gestão. I. Título.

CDD 621.38

This page was intentionally left blank

Contents

Acknowledgements iv

Abstract v

Resumo vi

List of Figures vii

List of Tables ix

Notation x

1 Introduction 1

1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.2 Device-to-Device Communication . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.3 Radio Resource Management (RRM) for Device-to-Device (D2D) Communication . 3

1.3.1 Peer Discovery and Pairing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.3.2 Mode Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.3.3 Resource Allocation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.3.4 Grouping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.3.5 Power Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

1.4 State of the Art . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

1.5 Thesis Organization and Contributions . . . . . . . . . . . . . . . . . . . . . . . . . 8

1.6 Scientific Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2 Methodology and System Modeling 11

2.1 Wireless System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

2.2 Radio Resource Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

2.3 Physical Resource . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

2.4 Multi-cell Scenario . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2.5 Wireless Channel Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

2.6 Transmission Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

2.7 Link-to-System Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

2.8 Imperfect Channel State Information . . . . . . . . . . . . . . . . . . . . . . . . . . 17

2.9 System Level Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

2.10Classification of Metrics Used in Energy Efficiency . . . . . . . . . . . . . . . . . . 18

i

2.10.1Energy Efficiency at the Network Level . . . . . . . . . . . . . . . . . . . . . 18

2.10.2Energy Efficiency at the System Level . . . . . . . . . . . . . . . . . . . . . . 19

2.10.3Energy Efficiency at the Component Level . . . . . . . . . . . . . . . . . . . 19

3 Energy Efficiency RRM Methods 20

3.1 Power Control (PC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

3.1.1 Equal Power Allocation (EPA) and Fixed Power . . . . . . . . . . . . . . . . 20

3.1.2 LTE Power Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

3.1.3 Soft Dropping Power Control (SDPC) . . . . . . . . . . . . . . . . . . . . . . 21

3.1.4 Closed Loop Soft Dropping (CLSD) . . . . . . . . . . . . . . . . . . . . . . . . 22

3.2 Downtilt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

3.2.1 Antenna Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

3.2.2 Electrical Antenna Downtilt . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

4 Results and Analysis 26

4.1 Power Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

4.1.1 Power Control Evaluation in a Micro-cell Scenario (Downlink) . . . . . . . 26

4.1.2 Power Control Evaluation in a Micro-cell Scenario (Uplink) . . . . . . . . . 30

4.1.2.1 LTE PC schemes and SDPC . . . . . . . . . . . . . . . . . . . . . . . 30

4.1.2.2 CLSD a hybrid PC scheme . . . . . . . . . . . . . . . . . . . . . . . 35

4.1.2.3 Impact of loads in PC schemes . . . . . . . . . . . . . . . . . . . . . 37

4.1.2.4 Imperfect Channel State Information (CSI) . . . . . . . . . . . . . . 38

4.1.2.5 Convergence of Soft Dropping (SD) . . . . . . . . . . . . . . . . . . 40

4.2 Antenna Downtilt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

4.2.1 Impact of Downtilt in a cellular network with D2D . . . . . . . . . . . . . . 41

4.2.2 SDPC in a Downtilt scenario . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

5 Conclusions 47

Appendix A Proof of convergence SDPC 49

Bibliography 53

ii

Acknowledgements

Initially, I thank God for having given me the strength to consolidate this dream. To

my parents, Wilson Nunes de Melo and Simone Cristina Lima de Melo, for their teachings

and moral values, my immense gratitude, my respect and my admiration, insurmountable in

words. My brother, Yago Willy Lima de Melo by support.

The professor Dr. Tarcisio Ferreira Maciel, my advisor, I thank for the support, incentive,

professional advice, and also of life. Thanks for shared knowledge, for all the help, the advice

that allowed me to complete my master’s thesis. Again, thank you for all of the help.

I also thank the members who accepted the invitation to participate M.Sc. defense boards:

Prof. Dr. Vicente Angelo de Souza Júnior, Prof. Dr. Emanuel Bezerra Rodrigues and Prof. Dr.

Francisco Rodrigo Porto Cavalcanti.

I would like to thank my colleagues from UFC 33, Rodrigo Batista, Carlos Filipe and José

Mairton for the discussions which greatly contributed for my growth as a researcher.

The postgraduate friends Daniel Araújo, Darlan Cavalcante, Diego Aguiar, Hugo Costa,

Igor Guerreiro, Igor Osterno, Juan Medeiros, Lázslon Costa, Marciel Barros, Marcio Caldas,

Paulo Garcia, Rafael Guimarães, Samuel Valduga, Victor Farias and Wilker Lima, with whom

I shared moments of tension, but mainly of much joy and laughter.

The GTEL, the research group in which I had the honorable opportunity to participate. It

was essential to make this work, contributing with computational resources, physical space

and, above all, rich intellectual environment. My thanks in particular to Ana Lívia, Isabel

Rabelo, Ogeniz Façanha and Vera.

The UFC by the infrastructure and the high standard of teachers who train professionals

capable of produce new knowledge and apply them to social reality. I would like also to

acknowledge CNPq for the scholarship support.

To all of you, my sincere appreciation and gratitude.

Yuri Victor Lima de Melo

”I have learned that a man only has the right to look down on another man

when it is to help him to stand up.”

Gabriel José García Márquez

Abstract

In a world where people count on their smartphone, smartwatch, tablet and other devices

to keep them connected wherever they go, they expect its application to run without problems,

such as dropped calls, slow download and choppy videos.

In this context, Device-to-Device (D2D) communication represents a promising technology,

because it is a direct and low-power communication between devices close, allowing to offload

the data transport network, increase spectral and power efficiency. From the subscriber point

of view, D2D means to use applications without problem and increase battery life. However, in

order to realize the potential gains of D2D communications, some key issues must be tackled,

because D2D communications may increase the co-channel interference and compromise the

link quality of cellular communications.

This master’s thesis focuses on Radio Resource Management (RRM) techniques, especially

Power Control (PC) schemes, to mitigate the co-channel interference for D2D communications

underlaying a Long Term Evolution (LTE) network, aiming at the reduction of the intra- and

inter- cell interference and at the improvement of energy efficiency. The main PC schemes

(e.g. OLPC, CLPC and SDPC) and a hybrid scheme (CLSD) are calibrated and used in macro- or

micro- multicell scenario, using different loads and imperfect Channel State Information (CSI).

In addition, the impact of downtilt is analyzed, which is used to adjust the coverage radius of

an Evolved Node B (eNB) and reduce co-channel interference by increasing cell isolation.

The numerical results indicate that PC schemes and downtilt, duly calibrated, can provide

gains to cellular and D2D communications. In other words, D2D technology can be used to

further increase the spectral and energy efficiency if RRM algorithms are used suitably.

Keywords: Device-to-Device (D2D) communication, Long Term Evolution (LTE)

network, Power Control (PC), Downtilt, Interference management, Energy efficiency

v

Resumo

Em um mundo onde as pessoas contam com smartphone, smartwatch, tablet e outros

dispositivos para mantê-las conectadas onde quer que vão, todos esperam que seus

aplicativos sejam executados sem problemas, tais como chamadas abandonadas, download

lento e vídeos com saltos.

Neste contexto, comunicação dispositivo-a-dispositivo (do inglês, Device-to-Device (D2D))

constitui uma tecnologia promissora, pois é um tipo de comunicação direta e utiliza baixa

potência entre dispositivos próximos, permitindo-se desviar o tráfego da rede móvel, aumentar

a eficiência espectral e de potência. Do ponto de vista do assinante, D2D significa usar

aplicação sem problemas e aumentar o tempo de vida da bateria do celular.

No entanto, a fim de realizar os ganhos potenciais das comunicações D2D, algumas

questões-chave devem ser abordadas, pois as comunicações D2D podem aumentar a

interferência co-canal e comprometer a qualidade do enlace das comunicações celulares.

Esta dissertação foca em técnicas de Gerenciamento de Recursos de Rádio (do inglês, Radio

Resource Management (RRM)) para mitigar a interferência co-canal para comunicações D2D

que se baseiam na Evolução de Longo Prazo (do inglês, Long Term Evolution (LTE)), visando

a redução da interferência intra- e inter-celular e na melhoria da eficiência energética. Os

principais esquemas de Controle de Potência (do inglês, Power Control (PC)) (e.g. OLPC,CLPC

e SDPC) e um esquema híbrido (CLSD) são calibrados e utilizados no cenário macro ou micro

multicelular, usando diferentes cargas e Informação do Estado do Canal (do inglês, Channel

State Information (CSI)) perfeita ou imperfeita. Além disso, o impacto da inclinação da antena

(downtilt) é analisado, que é usada para ajustar o raio de cobertura de uma Evolved Node

B (eNB) e reduzir a interferência co-canal, aumentando o isolamento de células.

Os resultados numéricos indicam que os regimes de controle de potência e inclinação

da antena, devidamente calibrados, podem fornecer ganhos para a comunicação celular e

D2D. Em outras palavras, a tecnologia D2D pode ser utilizada para aumentar ainda mais

a eficiência de espectro e a eficiência energética se algoritmos de RRM forem utilizados

adequadamente.

Palavras-chave: dispositivo-a-dispositivo (D2D), Redes LTE, Controle de potência

(PC), Inclinação da antena (Downtilt), Gerenciamento de interferência, Eficiência

energética

vi

List of Figures

1.1 RRM techniques for D2D communications . . . . . . . . . . . . . . . . . . . . . . . 3

1.2 RRM procedures in D2D generic scenario. . . . . . . . . . . . . . . . . . . . . . . . 6

2.1 Classification of wireless communication networks according to the coverage. . . 11

2.2 OFDMA frame structure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2.3 Coverage area of the multi-cell scenario. . . . . . . . . . . . . . . . . . . . . . . . . 14

2.4 Communication within a cell for both directions(Downlink (DL) and Uplink (UL)),

where the solid lines describe the interesting links and the dashed lines represent

the interfering links. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.5 Curves of link-level used for link adaptation. . . . . . . . . . . . . . . . . . . . . . 16

2.6 Imperfect CSI using feedback delay. . . . . . . . . . . . . . . . . . . . . . . . . . . 17

3.1 Target SINR as function of a variable transmit power. . . . . . . . . . . . . . . . . 21

3.2 Azimuth orientation and downtilt in a macrocell scenario. . . . . . . . . . . . . . 24

4.1 Calibration of SD algorithm regarding the total system spectral efficiency and

transmit power. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

4.2 SINR and interference power of cellular and D2D communications by applying

SD and EPA schemes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

4.3 System spectral efficiency by applying SD and/or EPA to cellular algorithms

and/or D2D transmitters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

4.4 Calibration of the SDPC scheme by applying it to cellular or D2D links. The

PC range ∆P = 0dB gives the performance of fixed power approach. Minimum

target Signal to Interference-plus-Noise Ratio (SINR) values are simulated until

Γmin = −5 dB because the SINR threshold of the lowest MCS is −6.2 dB. . . . . . . 30

4.5 Calibration of the OLPC scheme by applying it to cellular or D2D links. The

pathloss compensation factor α = 0 gives the No-PC performance. . . . . . . . . . 31

4.6 SINR and interference power levels by applying SDPC and OLPC schemes to

cellular or D2D links. No-PC and fixed power approaches are considered as

baselines. No-PC (cellular) represents the conventional scenario without D2D

communications underlaying the cellular network. . . . . . . . . . . . . . . . . . . 32

4.7 Total system spectral efficiency by applying OLPC to D2D links without PC for

cellular links (No-PC approach). The pathloss compensation factor α of the

OLPC scheme is varied for target Signal to Noise Ratio (SNR) Γk = 10 dB. The

conventional scenario considers the No-PC approach in its cellular links. . . . . 33

vii

4.8 SINR by applying power control schemes to cellular and D2D links. . . . . . . . . 34

4.9 Performance of PC schemes for cellular and D2D communications. . . . . . . . . 35

4.10Spectral efficiency of PC schemes for cellular and D2D communications. . . . . . 36

4.11Power efficiency of PC schemes for cellular and D2D communications. . . . . . . 36

4.12Total spectral efficiency comparison for different loads. . . . . . . . . . . . . . . . 37

4.13Power efficiency comparison for different loads in cellular and D2D

communications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

4.14Total spectral efficiency for different delays. . . . . . . . . . . . . . . . . . . . . . . 39

4.15Power efficiency for different delays in cellular and D2D communications. . . . . 40

4.16Detailed description of calculation of convergence. . . . . . . . . . . . . . . . . . . 40

4.17Convergence of Soft Dropping (SD). . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

4.18Behavior of spectral and power efficiency for different levels of tilt. . . . . . . . . 42

4.19SINR and interference levels by applying downtilt. . . . . . . . . . . . . . . . . . . 43

4.20System spectral efficiency of cellular and D2D communications in scenario with

and without downtilt. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

4.21Outage reduction for different levels of tilt. . . . . . . . . . . . . . . . . . . . . . . . 44

4.22Total spectral efficiency and power efficiency (SDPC in cellular and No-PC in D2D

links). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

4.23Total spectral efficiency (No-PC in cellular and SDPC in D2D links). . . . . . . . . 46

viii

List of Tables

2.1 Transmitter and receiver sets for D2D communications in both UL and DL

communication phases. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.2 SINR thresholds for link adaptation . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

2.3 Simulation parameters for urban-macrocell and microcell environments. . . . . . 18

2.4 Metrics Used in Energy Efficiency. . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

4.1 Relative gains of performance by applying the SD algorithm to cellular and D2D

communications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

4.2 Relative gains of system spectral efficiency by applying the SD algorithm to D2D

communications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

4.3 Relative gains by applying SDPC and OLPC to cellular or D2D links in

comparison to the No-PC approach (%). . . . . . . . . . . . . . . . . . . . . . . . . 31

4.4 Relative performance gains of Open Loop Power Control (OLPC) for D2D links

compared with no-PC for D2D links (%). . . . . . . . . . . . . . . . . . . . . . . . . 33

4.5 Calibration of σ for CLPC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

4.6 Closed Loop Soft Dropping (CLSD) parameters . . . . . . . . . . . . . . . . . . . . 35

4.7 Spectral efficiency relative gains applying CLSD compared with other PC

schemes (%). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

4.8 Power efficiency relative gains applying CLSD compared with other PC schemes

(%). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

4.9 Power efficiency relative gains for different downtilt angles compared without

downtilt (%). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

ix

Notation

Acronyms

3G 3rd Generation

3GPP 3rd Generation Partnership Project

4G 4th Generation

5G 5th Generation

BLER BLock Error Rate

BS Base Station

CDF Cumulative Distribution Function

CLPC Closed Loop Power Control

CLSD Closed Loop Soft Dropping

CPU Central Processing Unit

CoMP Coordinated Multi-Point

CSI Channel State Information

D2D Device-to-Device

DIST Distance-based Grouping

DL Downlink

eNB Evolved Node B

EPA Equal Power Allocation

GSM Global System for Mobile Communications

GPL GNU General Public License

IMT International Mobile Telecommunications

ITU International Telecommunication Union

LTE Long Term Evolution

LSI Large-Scale Integration

MCS Modulation and Coding Scheme

MIMO Multiple Input Multiple Output

MR Maximum Rate

MS Mode Selection

MSE Mean Squared Error

NLOS Non-Line of Sight

NMSE Normalized Mean Square Error

OFDMA Orthogonal Frequency Division Multiple Access

OFDM Orthogonal Frequency Division Multiplexing

OLPC Open Loop Power Control

x

PAIR D2D Pair Gain-based Grouping

PC Power Control

PRB Physical Resource Block

QAM Quadrature Amplitude Modulation

QoS Quality of Service

QoE Quality of Experience

QPSK Quadri-Phase Shift Keying

RAN Radio Access Network

RA Resource Allocation

RET Remote Electrical Tilt

RB Resource Block

RM Rate Maximization

RRA Radio Resource Allocation

RRM Radio Resource Management

RR Round Robin

SCM Spatial Channel Model

SD Soft Dropping

SDPC Soft Dropping Power Control

SINR Signal to Interference-plus-Noise Ratio

SIR Signal to Interference Ratio

SNR Signal to Noise Ratio

TTI Transmission Time Interval

UE User Equipment

UL Uplink

VET Variable Electrical Tilt

WCDMA Wideband Code Division Multiple Access

WINNER Wireless World Initiative New Radio

WPAN Wireless Personal Area Network

WWAN Wireless Wide Area Network

xi

Chapter 1Introduction

1.1 Motivation

The development of Radio Access Networks (RANs) has provided triple-play services

(i.e. voice, video and data) anytime and anywhere. These features demand high spectral

efficiency and so it is essential to ensure interoperability of radio access technologies and

convergence of different services. Therefore, the International Telecommunication Union (ITU)

has established a set of requirements for a high performance 4th Generation (4G) [1] of wireless

communication systems. The key requirements are:

◮ high quality mobile services,

◮ user equipment suitable for worldwide use,

◮ user-friendly applications, services and equipment,

◮ worldwide roaming capability,

◮ compatibility of services within International Mobile Telecommunications (IMT) and with

fixed networks.

Modern wireless networks e.g. 5th Generation (5G) need to be efficiently designed in order

to support as much calls, data transmissions and mobile services as possible, and still extend

the battery lifetime of User Equipments (UEs). Additionally, the successful operation of

modern telecommunication systems is dependent, in part, on sophisticated real-time control

mechanisms. Energy-efficient wireless networks have become an important research topic

in the last years due to the increasing rate of data traffic and the quick growing of energy

consumption [2].

Device-to-Device (D2D) communications underlaying a cellular network can improve

resource utilization and potentially lead to a reduced power consumption. However, D2D

communications may increase the co-channel interference and compromise the link quality of

cellular communications [3]. Power Control (PC) is an important Radio Resource Management

(RRM) functionality in wireless communication systems, which adapts the transmit power of

the communicating devices to ensure a target Quality of Service (QoS) level, thus limiting

interference and prolong the battery lifetime. Therefore, the design of PC schemes becomes

attractive in order to keep interference under control, protect cellular communications, and

get energy-efficient transmissions [4].

Another technique used to keep interference under control is called antenna downtilt,

which is responsible for changing the antenna radiation pattern. By utilizing antenna

1.2. Device-to-Device Communication 2

downtilt, signal level within a cell can be improved and interference radiation towards other

cells can be effectively reduced due to the antenna radiation pattern. However, an excessively

large downtilt angle might lead to coverage problems at cell border areas. Therefore, it is vital

to define a suitable downtilt angle [5].

1.2 Device-to-Device Communication

D2D is an attractive means of expanding mobile network capacity, user experience, energy

efficiency and corverage. When a direct communication occurs between two UEs, this

communication is called D2D. To provide high spectral efficiency, advanced techniques are

needed to manage and control interference, because D2D users are added to cell.

D2D communication was first cited in [6] to allow multihop relays in cellular networks.

Then others papers were published [7, 8], where the main investigated subject was the

potential of D2D communications for improving spectral efficiency of cellular networks. Other

potential benefits of incorporating a D2D communication in a cellular network is metioned

such as peer-to-peer communication [9, 10], video dissemination [11], machine-to-machine

(M2M) communication [12,13] and cellular offloading [14,15].

As mentioned above, D2D communications are added to the cells of a cellular sytem. Thus,

it is important to understand how the resources are divided in the cellular network when D2D

communication is used. D2D communication can occur on unlicensed spectrum (outband)

or on cellular spectrum (inband). The inband technique can be subdivided in a group where

all spectrum of cellular network can be used to cellular or D2D communication and another

group where each communication uses a specific portion of the spectrum. These techniques

are called, respectively, underlay and overlay.

There are works about inband and outband D2D communication in the literature [16,17].

When D2D communication is used outband, the main problems are related to coordinating

the communication over different bands due a second radio interface (e.g., WiFi Direct [16]

and bluetooth [17]). Regarding the underlay case, until now, the main question is the problem

of interference mitigation between D2D and cellular communications [18–20].

The features of the D2D are described below [16,20]:

◮ Unlicensed spectrum (outband): WiFi and Bluetooth operate in unlicensed spectrum,

without any centralised control of usage or interference. This is not generally a problem

when usage densities are low, but it would become a major limitation as proximity-based

services proliferate. Throughput, range and reliability would all suffer;

◮ Security: The security features of WiFi and Bluetooth are much less robust than those

used in public cellular systems. They would not be adequate for major public services

and they would be unsuitable for public safety applications;

◮ Radio resource management: Unlike Bluetooth and WiFi, Long Term Evolution (LTE)

operates in licensed spectrum and the radio resources are carefully managed by the

network, to minimise interference and maximise the performance of the system. The

same mechanisms can be extended to D2D;

◮ Performance: Direct communication between nearby devices may be able to achieve

even higher throughput and lower latency than communication through an LTE base

station. For example, the devices may be closer to each other than either of them is to

the nearest base station and a busy base station may be a bottleneck. The network can

1.3. RRM for D2D Communication 3

still exert control over the radio resources used for these connections, to maximise the

range, throughput and overall system capacity;

◮ Spectrum reuse: D2D could enable even tighter reuse of spectrum than can be achieved

by LTE small cells, by confining radio transmissions to the point-to-point connection

between two devices;

◮ Network load: Relieving the base stations and other network components of an LTE

network of some of their traffic-carrying responsibilities, for example carrying rich media

content directly between mobile terminals, will reduce the network load and increase its

effective capacity;

◮ Energy efficiency: Integrating D2D into the LTE system provides the opportunity to

achieve energy-efficient device discovery, for example by avoiding the need to scan

for other wireless technologies, by synchronising the transmission and reception of

discovery signals to minimise their duty cycle and by waking application software only

when relevant devices are found in the local area. Meanwhile, direct transmission

between nearby devices can be achieved with low transmission power.

1.3 RRM for D2D Communication

New challenges related with interference appear when D2D communications use cellular

spectrum. Thus, it is necessary to improve and create RRM techniques for D2D

communications such as mode selection, user grouping, Power Control and adaptive

scheduling. This section gives an explanation about RRM for D2D communications, which

are illustrated in Figure 1.1.

Begin

Peer Discovery

Set D2D Pairs

Band Selection

End

Link Establishment

Power Control

Precoding Filters

Update Neighbors

List

Grouping / Scheduling

Mode Selection

RRM for D2D Communications

Figure 1.1: RRM techniques for D2D communications

1.3.1 Peer Discovery and Pairing

The goal of peer discovery is to look for UEs, which are candidates to establish D2D links.

A UE or Evolved Node B (eNB) can be responsible for this procedure, in which they transmit

beacon signals to identify its neighbors and analyze their beacon intensity. UEs have a

higher probability to become candidate for D2D communication when their beacon intensity

is strong. After the first step, the next step is the device pairing. Pairing is responsible for

determining which D2D candidates establish D2D links in the cellular network [21]. The first

and second step are well known in, for example, Bluetooth, where master node identifies in a

range the devices wishing to participate of a specific connection [22].

1.3. RRM for D2D Communication 4

1.3.2 Mode Selection

Mode Selection (MS) is an important RRM technique in D2D communication, which

determines whether UEs can communicate directly or not (i.e. via the Base Station (BS)). In

[23], the author proposes a mode selection procedure that takes into account the link quality

of the D2D link and the different interference situation when sharing cellular Uplink (UL)

or Downlink (DL) resources. The results ensure a reliable D2D communication with limited

interference to the cellular network.

Other study about MS is realized in [24]. Therein, the author derives the system

equations that can be used to analyze a system where both cellular communication and D2D

communication can share the same resources. Via numerical analysis it was shown that

communication mode selection needs to be designed carefully to prevent deteriorating the

system performance. The results show that the main affecting factors for the performance

gain from D2D are local communication probability and maximum distance between

communicating devices. In other words, D2D communication is propitious when the UE

is close to BS and the distance between the UEs of a D2D pair is short.

1.3.3 Resource Allocation

The purpose of the Resource Allocation (RA) is to select Physical Resource Block (PRB)

of a set of available PRB for each transmission/reception in cellular or D2D communication.

In [25], the authors use a method with which D2D communications can reuse the resources of

more than one cellular user. The authors assume that PRBs can be selected with an optimal

resource allocation method using the Channel State Information (CSI) of all involved links.

The results show that the proposed method is the optimal method when D2D are located in

most parts of the cell area and the method achieves better performance when the D2D pair

becomes closer to the cell edge.

The Round Robin (RR) and Maximum Rate (MR) scheduling algorithms are well-known in

literature and they can be used in D2D communication [26]. The principle of RR algorithm is

to be resource fair with each user. It is accomplished by assigning the same number of PRBs

to every user. The principle of Rate Maximization (RM) algorithm is to assign resources to the

users which maximize system rate. The algorithm is performed for each PRB and resource are

assigned the users with the largest channel gain on that PRB. The results showed that higher

throughput gains are achieved when scheduling prioritizes the D2D mode due to proximity.

1.3.4 Grouping

The grouping is the key technique to achieve high reuse gains, because it is used to choose

which cellular and D2D links should share a PRB. For example, it is possible to choose

cellular and D2D users to share a resource randomly and, therefore, no channel information

is used.

In [27], the authors developed several grouping algorithms. The Distance-based Grouping

(DIST) algorithm’s basic idea is to group the D2D transmitters that are farthest from the eNB

with the scheduled cellular UEs as to obtain resource reuse gains without much losses to

cellular communications performance. Therein, the D2D Pair Gain-based Grouping (PAIR)

method is founded on the fact that the proximity between D2DTx and D2DRx is an important

parameter for D2D communications and this aspect is prioritized for resource sharing. It is

assumed that the large-scale fading gain for the link between these nodes is made available

to the eNB, which uses it to represent the effective radio distance between nodes. This gain

can be estimated and reported to the eNB by the D2DTx, D2DRx or both.

1.4. State of the Art 5

1.3.5 Power Control

Improper use of transmit power can harm all previous blocks of RRM, because a high

transmit power for a cellular or D2D transmitter can increase the interference level of system,

decrease the QoS and reduce battery life. So, it is clear that PC schemes are important in

traditional cellular network and become essential when D2D links are added to the network.

The D2D links have to adjust their transmit powers seeking to increase spectral and power

efficiency, while cellular links keep a acceptable QoS. In [28], a dynamic PC mechanism

is proposed to reduce interference generated by D2D communications and improve the

performance of cellular communications in DL. The proposed algorithm has two phases.

In the first phase, the eNBs assigns resources to D2D communications by reusing the same

resources allocated to cellular UEs. Then, PC is usually applied for D2D communications to

decrease interference to cellular UEs. For this goal, the eNB adjusts the transmit power of the

D2D transmitter based on estimated channels gains between each desired link.

In [29], the authors consider that a cellular UE needs to communicate with an eNB in UL

while multiple D2D links coexist in the common spectrum. Two forms of PC were proposed:

centralized and distributed. Centralized PC occurs when D2D links are managed by eNB.

In this case, the eNB needs to know the global CSI. The proposed distributed PC sets the

transmit power of D2D-capable UEs based on the knowledge of direct link information and

the minimum channel gain that is fixed and known by all UEs.

1.4 State of the Art

D2D communication is a technology used to improve QoS and Quality of Experience (QoE)

of the users, while it provides the increase of resource utilization in cellular networks,

because it can operate in licensed and unlicensed spectrum bands. In other words,

D2D communications can underlay a cellular network, employing the same radio resource

to improve the system efficiency. In a cellular network, where UE have traditional

communications via eNB in a LTE system, UEs have the capacity to create a direct

communication with each other over D2D links.

However, it is necessary to manage all these links and, for this purpose, the eNB becomes

responsible for controlling the radio resources and set transmission parameters, such as,

communication duration and transmit power.

There are several industrial and academic researches related with D2D communications,

which show and explain the benefits of D2D communications to the next-generation of cellular

networks, such as:

◮ provide a better energy efficiency;

◮ offload cellular networks;

◮ improve system capacity;

◮ increase coverage;

◮ improve QoS and QoE.

The combination of cellular and D2D communications opens several issues such as the

analysis and design of techniques related to optimization, signal processing, decision theory

and layer perspectives (e.g. physical, MAC, network and application). Figure 1.2 illustrates

the main RRM procedures in a generic scenario with an eNB surrounded by UEs using cellular

and D2D communications.

1.4. State of the Art 6

Mode Selection

Peer Discovery

(Beacon)

D2D Mode

Cellular M

ode

Cellular ModeCellular vs D2D

Interference

Management

P ee r

D

isco

ver y

(B

eaco

n)

Peer Discovery

(Beacon)

Peer Discovery List

HotspotHotspot

Figure 1.2: RRM procedures in D2D generic scenario.

Peer discovery in cellular networks has been studied in [30], where the authors propose

a synchronous device discovery solution for networks based on the observations of the time

synchronization. The results indicated that the solution has a large advantage over WiFi for

device discovery, both in terms of range and energy efficiency. In [9], the authors focus on

peer discovery for D2D communication in LTE networks. A new distributed discovery protocol

is proposed for UEs to broadcast their presence. In the proposed protocol, UEs transmit

beacons periodically to advertise their presence. The purpose of such control is for an eNB

to minimize the required Resource Blocks (RBs) for beacon transmission, while still providing

efficient peer discovery for D2D UEs. The authors concluded that the algorithm provides a

good performance in discovery of UEs with mobility in LTE networks.

Regarding the resource assignment between cellular and D2D users, in [31], a heuristic

algorithm considering channel gain information appropriately selects the shared radio

resources for both users. In [32], the authors use the diversity in the cellular network to

improve the network capacity. In [33], the system spectral efficiency is increased by allowing

D2D users to reuse the resources of more than one cellular user in a system where perfect

CSI is assumed.

Mode selection has been studied in [34–36]. In [34] semi-analytical studies have shown

that when D2D communications share the same resources as the cellular network, significant

gains in total throughput can be achieved compared to the conventional case, namely by

the jointly and optimal allocation. However, numerical analyses have also shown that

mode selection algorithms need to be designed carefully in order to prevent deteriorating

the whole system performance. In [35], the authors derive equations that capture the

network information such as link gains, noise levels, and Signal to Interference-plus-Noise

Ratios (SINRs). The results shown that the main factors affecting the performance gain

of D2D communication are the local communication probability and maximum distance

between communicating nodes, as well as the mode selection algorithm. In [36], the eNB

can decide whether the underlaying D2D pair should reuse cellular resources, get dedicated

resources or communicate via eNB. One conclusion drawn from this paper is that an

optimal communication mode selection strategy does not only depend on the quality of the

link between D2D terminals and the quality of the link towards the eNB, but also on the

interference situation.

PC schemes are one of keys to the harmonious coexistence between cellular and D2D

communications. In this context, the transmit power of both communications need to be

adjusted by the eNB based on channel gain, QoS demands, coverage and/or target SINR.

A PC method for D2D communications was proposed in [37] to maximize the network sum

rate. Its optimality is discussed under practical constraints such as minimum and maximum

spectral efficiency, and maximum transmit power. In [38], a power minimization solution

1.4. State of the Art 7

with joint subcarrier allocation, adaptive modulation, and mode selection was proposed to

guarantee the QoS demand of D2D and cellular communications. A simple PC scheme

was proposed in [39] to regulate the transmit power of D2D-capable UEs and protect the

existing cellular links in a single-cell scenario and deterministic network model. The algorithm

imposes constraints on the SINR to allow quality degradation of cellular links until fixed levels

are reached in DL and UL communication phases.

Different UL PC schemes have been studied for D2D communications in the literature [40,

41], including fixed transmit power schemes, fixed target Signal to Noise Ratio (SNR) schemes,

and LTE PC schemes – Open Loop Power Control (OLPC) and Closed Loop Power Control

(CLPC). In [40], a new PC scheme with double thresholding that coordinates the transmit

power of D2D and cellular UEs to maximize the cell throughput and guarantee QoS levels

is proposed in a scenario composed of a cellular UE and a D2D pair. The results show a

throughput improvement in comparison with LTE OLPC. In [41], the authors use the LTE

OLPC for cellular links and study other PC schemes for D2D links. The authors conclude

that the LTE PC schemes gets close (especially for the conventional cellular UEs) in terms

of transmit power and SINR levels to an optimization-based approach aiming to increase

spectrum usage efficiency and to reduce sum power consumption.

In [42], the Soft Dropping Power Control (SDPC) scheme adjusts the transmit power to meet

a variable target SINR in an UL single-carrier system. In [43], the SDPC scheme was used

to protect cellular and D2D communications from mutual interference in a DL Orthogonal

Frequency Division Multiple Access (OFDMA) system. It improved the spectral efficiency of

cellular UEs in 14% and still significantly reduced the power of D2D transmitters in 49%

without harming the spectral efficiency achieved by D2D receivers. Thus, the SDPC scheme

appears as a promising solution to protect cellular UEs from the interference caused by D2D

communications.

In [44], the authors examine the consequence of antenna downtilt and UL PC on the system

level performance considering a realistic multicell 3D channel model. A highlight of the paper

is the performance evaluation considering different downtilt angles and OLPC. The paper

shows that angles between 4° and 8° are good for cells with radius in the range of 300m based

on a urban-macro path loss model based on the WINNER II [45] channel model.

A study about the relation between load balancing and antenna tilt adjustment schemes is

one of the main contributions in [46]. The authors simulated different load balancing methods

based on combinations of cell association algorithms and antenna tilt. The potential gain of

traffic load balancing in terms of cell edge user throughput and significant cell edge user

throughput improvements were observed by the authors, in contrast to the fixed case. In [47],

antenna tilt adaptation was used to redistribute cell load from high congested areas to the

areas with less congestion by using the Simulated Annealing [48] meta-heuristic and lead to

efficient utilization of radio resources.

A research in field trial is detailed in [49]. The paper presents a set of UE locations, where

downtilt could increase Signal to Interference Ratio (SIR) by about 5 dB to 10 dB. Furthermore,

the effect of downtilt on the multi-path channel, location of the user and the eNB power is

investigated.

Strategies that exploit system-level analyses for the performance gains achieved with

Radio Resource Allocation (RRA) strategies for rate maximization in downlink multi-antenna

Coordinated Multi-Point (CoMP) systems are investigated in [50]. The authors realized

analysis of the antenna downtilt to mitigate inter-cell interference and concluded that spatial

diversity-based transmission schemes combined with downtilt provided satisfactory gains,

1.5. Thesis Organization and Contributions 8

especially for low loads expressed in number of active UEs per cell. In [51], antenna

tilt adaptation is used for capacity optimization using techniques to identify the dominant

interfering cells. Results show that the proposed technique identifies a reduced set of

potentially significant interfering cells among the neighbors which have considerable impact

on system performance.

In order to present basic effects on network coverage and capacity due to changes

in the antenna downtilt angle configuration when mechanical or electrical adjustment of

the downtilt is used, the paper [52] shows the percentage of covered area under certain

circ*mstances. The electrical adjustment of the downtilt angle performed slightly better than

the mechanical one. According to the results presented therein, the smaller the cell size the

larger the antenna downtilt should be; and the higher the traffic load per cell the smaller

the antenna downtilt should be. In [53], the potential gain of tilt optimization due to user

traffic distribution is investigated for the 3rd Generation Partnership Project (3GPP) urban

propagation environment. Therein, a traffic hotspot situation is assumed, the tilt of each

sector is adapted, and user throughput performance targets are defined. According to the

authors, the performance gain is larger for higher traffic densities at the hotspot.

1.5 Thesis Organization and Contributions

This thesis is organized as follows. In Chapter 2, we concentrate on the methodology

and system model that are applicable in cellular networks integrating D2D communications.

More specifically, we show the RRM for cellular communications and discuss about mode

selection, resource allocation, grouping and power control for D2D communications. The

benefits of D2D communications underlaying cellular networks are detailed in different topics

as security, performance and energy efficiency. Subsequently the details about physical

radio resources, wireless channel, transmission, link-to-system interface and imperfect CSI

modeling are addressed. Finally, we show the classification of metrics used to quantify energy

efficiency at network, system and component levels.

In Chapter 3, we explain about the efficiency energy methods used to analyze the D2D

scenarios addressed in Chapter 2. In this chapter we focus on baselines such as Equal Power

Allocation (EPA), Fixed Power and Fixed SINR, which are used to compare the efficiency of

Open Loop Power Control (OLPC), Closed Loop Power Control (CLPC), Soft Dropping Power

Control (SDPC) and Closed Loop Soft Dropping (CLSD). Next, we describe the formulation of a

simple and efficient downtilting, which is used to reduce undesired effects as inter- and intra-

cell interference.

In Chapter 4, we show the results of the performance evaluation of the referred PC schemes

in a macro-cell and in a micro-cell scenario using UL or DL bands. The main contributions

are:

◮ Show the performance of PC with variable target SINR levels in a multi-cell scenario,

◮ Compare the LTE PC schemes,

◮ Suggest and analyze the parameters for the CLPC scheme,

◮ Show the performance of PC with variable target SINR levels in a multi-cell scenario,

◮ Show the minimum performance impact on cellular communications for enabling D2D

gains in a multi-cell scenario,

◮ Propose an SDPC-like alternative as PC scheme,

1.6. Scientific Production 9

◮ Calibrate operating points of the considered PC schemes for energy efficiency of cellular

and D2D communications,

◮ Create and analyze the performance of CLSD,

◮ Test the performance of PC for different loads,

◮ Examine the impact of imperfect CSI,

◮ Show the convergence of SDPC,

◮ Implement the downtilt in the OFDMA system with D2D communications underlying

cellular networks,

◮ Show the impact of downtilt in a multi-cell scenario,

◮ Determine range of downtilt angles that impact positively on cellular and D2D

communications.

In Chapter 5, we summarize the main conclusions obtained along the master’s thesis.

Furthermore, we point out the main research directions that can be considered as extension

of the study performed in this master’s thesis.

1.6 Scientific Production

The contents and contributions present in this thesis were published and submitted with

the following information:

◮ Melo, Y.V.L; Batista, Rodrigo L.; Maciel, Tarcisio F.; Silva, Carlos F.M.e; da Silva, Jose

Mairton B.; Cavalcanti, Francisco R.P., “Power control with variable target SINR for D2D

communications underlying cellular networks,” in European Wireless 2014 (EW2014),

Barcelona, Spain, May 2014.

◮ Melo, Y.V.L; Batista, Rodrigo L.; Silva, Carlos F.M.e; Maciel, Tarcisio F.; da Silva, Jose

Mairton B.; Cavalcanti, Francisco R.P., “Power Control Schemes for Energy Efficiency

of Cellular and Device-and-Device Communications,” in Wireless Communications and

Networking Conference (WCNC), New Orleans, United State of America, March 2015.

◮ Melo, Y.V.L; Batista, Rodrigo L.; Silva, Carlos F.M.e; Maciel, Tarcisio F.; da Silva,

Jose Mairton B.; Cavalcanti, Francisco R.P., “Uplink Power control with variable target

SINR for D2D communications underlying cellular networks,” in Vehicular Technology

Conference (VTC2015-Spring), Glasgow, Scotland, May 2015.

In parallel to the work developed during the master’s course, I have been working on other

research projects, which are in the context of power allocation and grouping:

◮ Melo, Y.V.L; Rodrigues, E.B.; Lima, F.R.M.; Maciel, Tarcisio F.; Cavalcanti, Francisco

R.P., “Evaluation of Utility-Based Adaptive Resource and Power Allocation for Real

Time Services in OFDMA Systems,” in International Telecommunications Symposium

(ITS-2014), São Paulo, Brazil, August 2014.

◮ da Silva, Jose Mairton B.; Maciel, Tarcisio F.; C. F. M. e Silva, Batista,

Rodrigo L. and Melo, Y.V.L, “User Equipment Grouping for Device-to-Device

Communications Underlying a Multi-Cell Wireless System” in EURASIP Journal on

Wireless Communications and Networking (submitted).

1.6. Scientific Production 10

In the context of the same project, I have participated on the following technical reports:

◮ Silva, Carlos F.M.e; J. Mairton B. da Silva Jr.;Melo, Y.V.L; Maciel, Tarcisio F.; and

Cavalcanti, Francisco R.P., “RRM and QoS Management for 5th Generation Wireless

Systems”, GTEL-UFC-Ericsson UFC.40, Tech. Rep., March. 2015, First Technical

Report.

◮ Batista, Rodrigo L.; Silva, Carlos F.M.e; da Silva, Jose Mairton B.; Melo, Y.V.L;

Maciel, Tarcisio F.; and Cavalcanti, Francisco R.P., “Network-Assisted Device-to-Device

Communications”, GTEL-UFC-Ericsson UFC.33, Tech. Rep., Aug. 2014, Fourth

Technical Report.

◮ Batista, Rodrigo L.; Silva, Carlos F.M.e; da Silva, Jose Mairton B.; Melo, Y.V.L;

Maciel, Tarcisio F.; and F. R. P. Cavalcanti, “Network-Assisted Device-to-Device

Communications”, GTEL-UFC-Ericsson UFC.33, Tech. Rep., Jan. 2014, Third Technical

Report.

◮ Batista, Rodrigo L.; Silva, Carlos F.M.e; da Silva, Jose Mairton B.; Melo, Y.V.L;

Maciel, Tarcisio F.; and Cavalcanti, Francisco R.P., “Network-Assisted Device-to-Device

Communications”, GTEL-UFC-Ericsson UFC.33, Tech. Rep., Aug. 2013, Second

Technical Report.

◮ Rodrigues, E.B.; Lima, F.R.M.;Melo, Y.V.L; Costa Neto, Francisco Hugo; Maciel, Tarcisio

F.; and Cavalcanti, Francisco R.P., “Analysis and Control of Trade-Offs Involving QoS

Provision”, GTEL-UFC-Ericsson UFC.33, Tech. Rep., Aug. 2013, Second Technical

Report.

◮ Batista, Rodrigo L.; Silva, Carlos F.M.e; da Silva, Jose Mairton B.; Melo, Y.V.L;

Maciel, Tarcisio F.; and Cavalcanti, Francisco R.P., “Network-Assisted Device-to-Device

Communications”, GTEL-UFC-Ericsson UFC.33, Tech. Rep., Feb. 2013, First Technical

Report.

Chapter 2Methodology and System Modeling

This chapter covers the fundamental issues about the methodology, system modeling and

features of Device-to-Device (D2D) communications, so that a reader without prior knowledge

could understand the problems and challenges in such systems. Terminology related to the

scope of this master’s thesis are presented in more detail. The remainder of this chapter report

is structured as follows. In Sections 2.1 and 2.2 described basic features of wireless network

and traditional Radio Resource Management (RRM) are presented. In the Sections 1.2 and

1.3, detailed aspects of D2D communication and RRM. In Sections 2.3, 2.4, 2.6 and 2.7 are

detailed the system model. Finally, Sections 2.8, 2.9, 2.10 presents imperfect Channel State

Information (CSI), simulation parameters and classification of metrics for energy efficiency.

2.1 Wireless System

The traditional standards of wireless communication can be classified in terms of coverage,

as shown in Figure 2.1. Wireless Personal Area Network (WPAN) is used in personal networks

(i.e. at small coverage) while Wireless Wide Area Network (WWAN) can cover several kilometers

and provide service to thousands of users.

Figure 2.1: Classification of wireless communication networks according to the coverage.

After the success of the Global System for Mobile Communications (GSM), new researches

have been conducted by academy and industry to improve general aspects as Quality of

Experience (QoE), security and cost, in addition to specific aspects as spectral efficiency,

power efficiency and new communication architectures.

The goals of the 3rd Generation (3G) of wireless communication systems were

announced by International Telecommunication Union (ITU) and called International Mobile

Telecommunications (IMT)-2000. By request of the ITU, several organizations joined the

3rd Generation Partnership Project (3GPP) and described its ideas for 3G networks. The

2.2. Radio Resource Management 12

outcome of the discussions was sent to ITU, which was responsible for the choice and

documentation of the proposed system. The system approved as 3G should provide worldwide

roaming, high transmissions rates (e.g. minimum of 2Mbit/s to low mobility users and

348Kbit/s to high mobility) [54].

2.2 Radio Resource Management

The goal of a communication company is to provide a capable network to keep the

maximum amount of clients with a determined Quality of Service (QoS) level. To ensure a

minimum QoS level is necessary to overcome several challenges (e.g. propagation, traffic and

interference) present in the cellular communication environment. In order to ensure high data

rate, coverage and satisfactory QoS it is fundamental to apply RRM, in other words, RRM is a

set of techniques, which ensure system capacity while the requirements of coverage and QoS

of the users are satisfied, overcoming difficulties inherent to radio propagation.

The traditional RRM techniques can be grouped into three categories: Power Control (PC),

mobility control (handover) and congestion control. PC is very important in systems that

employ frequency reuse, such as in Wideband Code Division Multiple Access (WCDMA). In

this case, all users use the same frequency and, therefore, it is important to have an efficient

interference control. Thus, PC chooses the lowest transmit power necessary to achieve a target

QoS level, otherwise a user poorly managed (in terms of transmit power) can harm links of

all users in system [55]. The mobility control is necessary when an user changes its location.

The system must provide the switching of all radio resources from one cell to another, so that

the user does not suffers any harm in his/her QoS [56].

The congestion control can be subdivided into admission control, load control and

scheduling. In congestion control, admission control and load control work together to offer

stability of QoS, coverage and capacity. There are strategies to block the access of new users

or make handover to balance the load, while keeping the stability of the system. In others

words, the admission control decides if a new connection must be established or not, while

load control tries to keep active communications at an acceptable QoS level by interrupting

(bad) connections in progress or performing handover. Finally, scheduling is responsible

for exploring the physical resources available (e.g. time, frequency and code) in an effort to

achieve fairness and capacity in the system [57].

2.3 Physical Resource

For Long Term Evolution (LTE), 3GPP specifies the Orthogonal Frequency Division Multiple

Access (OFDMA) technology as radio access technique. OFDMA allows to exploit frequency

and multiuser diversities, since different subcarriers present different fading if sufficiently

apart and channel fading also varies for User Equipments (UEs) at different locations. Thus,

one can allocate subcarriers to UEs depending on their channel fading state/channel quality.

As it is well-known, OFDMA is based on Orthogonal Frequency Division Multiplexing (OFDM)

and enables the transmission of multiple parallel low-rate data streams over orthogonal

subcarriers, which correspond to narrow band channels created by sub-dividing the system

bandwidth. It allows each UE to be assigned resources that are orthogonal in time and

frequency. Usually, due to signaling constraints, subcarriers are not allocated individually,

but in blocks of adjacent OFDM subcarriers, which represent the Physical Resource Blocks

(PRBs) [58]. Channel coherence bandwidth is assumed larger than the bandwidth of a PRB

leading to flat fading over each PRB. For a given PRB, the complex channel coefficients

considered in this thesis correspond to those associated with the middle subcarrier of the

2.4. Multi-cell Scenario 13

considered PRBs. In 3GPP LTE, an OFDM frame structure takes the form of a frequency-time

resource grid as shown in Figure 2.2.

Sub-frame 1 Sub-frame 2 Sub-frame T

Figure 2.2: OFDMA frame structure.

As it is seen in Figure 2.2, the bandwidth has NPRB PRBs and the Transmission Time

Intervals (TTIs) are grouped into frames, each composed of NSUBFRAME subframes, where each

subframe supports Downlink (DL) or Uplink (UL) links and takes the duration of one TTI. The

PRB is defined as one subframe in the time domain, which is divided into 14 symbols, and

12 contiguous OFDM sub-carriers spaced of 15 kHz in the frequency domain. The minimum

allocable resource in LTE systems is the PRB. This unit corresponds to the available resource

that can be assigned to UEs by an Radio Resource Allocation (RRA) function of the system.

2.4 Multi-cell Scenario

The multi-cell scenario considered in this master’s thesis corresponds to a cellular network

with Evolved Node Bs (eNBs) uniformly distributed over the coverage area. It was assumed

that each eNB is placed at the center of a cell site, which is represented by a regular

hexagon. Two 3GPP fading environments are considered: urban-microcell and macrocell [59].

Graphically, the multi-cell scenario is shown in Figure 2.3 for both 3GPP environments.

As depicted in Figure 2.3, in the urban-microcell environment the site comprises only

a single cell while in the urban-macrocell environment it comprises three cells. In the

considered notation, it is assumed that the multi-cell scenario is composed of NCELL cells

and serves a number NUE of UEs uniformly distributed over its coverage area. Each UE is

equipped with NUE-ANT omnidirectional antennas.

Each cell comprises a hotspot zone located near the cell-edge in order to model situations

in which D2D communications are likely to happen [3]. Herein, 50% of the total number of

UEs within the cell are clustered inside a 50× 120m hotspot zone while the remaining UEs are

uniformly distributed over the cell area. Considering that UEs inside the hotspot are close

to each other and far from most cellular UEs, pairs of D2D-capable UEs are obtained by a

simply random pairing procedure [3]. Figures 2.4(a) and 2.4(b) exemplify cellular and D2D

communications in such hotspot zones in the urban-microcell environment for the DL/UL

communication phase

2.4. Multi-cell Scenario 14

eNB

site

(a) Urban-microcell environment.

eNB

3-cell site

(b) Urban-macrocell environment.

Figure 2.3: Coverage area of the multi-cell scenario.

eNB

UE2

UE3

UE1

Ho

tsp

ot

D2D Rx

D2D Tx

Cellular UE

D2D link

Interfering links

Cellular link

(a) Cellular and D2D communication in DL.

eNB

UE2

UE3

UE1

Ho

tsp

ot

D2D Tx

D2D Rx

Cellular UE

D2D link

Interfering links

Cellular link

(b) Cellular and D2D communication in UL.

Figure 2.4: Communication within a cell for both directions(DL and UL), where the solid lines describethe interesting links and the dashed lines represent the interfering links.

Due to D2D communication, in both UL and DL communication phases both UEs and

cells may be transmitters or receivers at the same TTI. Let T denote the transmitters set

and R the receivers set. In a given communication phase, one can include the set of all

cells, denoted by C = {CELL1,CELL2, . . . ,CELLNCELL}, and/or the set of all UEs, denoted by

U = {UE1,UE2, . . . ,UENUE}, in the multi-cell system. In Table 2.1, transmitter and receiver

sets are summarized for both UL and DL communication phases.

Table 2.1: Transmitter and receiver sets for D2D communications in both UL and DL communicationphases.

Parameter DL UL

Transmitters set (T ) C ∪ U U

Receivers set (R) U C ∪ U

Number of transmitters (NTX) NCELL +NUE NUE

Number of receivers (NRX) NUE NCELL +NUE

It is assumed that frequency resources can be fully reused in all cells. Since the number

of UEs is typically larger than the number of available resources, UEs have to be scheduled

by the RRA algorithms. As shown in Figure 2.2, in each subframe there exist NPRB PRBs in

the system and each of them might be assigned to one or more UEs in each cell.

2.5. Wireless Channel Model 15

2.5 Wireless Channel Model

The modeling of the complex channel coefficients includes propagation effects on the

wireless channel, namely, path loss, shadowing, short-term fading and also includes the

antenna gains. The distance dependent Non-Line of Sight (NLOS) pathloss in the microcell

environment is based on the COST 231 Walfish-Ikegami NLOS model, whereas the pathloss

in the macrocell environment is based on the modified COST 231 Hata urban propagation

model. Particular aspects of path loss modeling for both urban-macrocell and urban-microcell

environments are described in [59]. Path loss model for macrocell and microcell environments

are 34.5 + 35 log10(d) and 35.7 + 38 log10(d), respectively. Slow channel variations due to

shadowing are modeled by a lognormal distribution of zero mean and standard deviation

σsh. For D2D communications, while the large-scale shadowing is defined according to

environment, the path loss model [60] employed for both environments is given by

PL(d) = 37 + 30 log10(d). (2.1)

Concerning the small-scale fading, the Spatial Channel Model (SCM) is considered. SCM is

a stochastic channel model developed by 3GPP for evaluating Multiple Input Multiple Output

(MIMO) system performance and incorporates important parameters such as phases, delays,

Doppler frequency, and ray angles [59]. The spatial characteristics of the SCM are described

by scatterers and clusters of scatterers placed over the considered scenario. Details of relevant

parameters for the SCM as well as their values are addressed in [59]. In this master’s thesis,

the SCM simulator available in [45]1 is used for obtaining the small-scale fading, which is in

accordance with the SCM specified in [59].

2.6 Transmission Model

It is necessary to calculate the Signal to Interference-plus-Noise Ratio (SINR) in both UL

and DL communication for each receiver in order to estimate data rates. When considering

the transmissions on a single PRB of the multi-cell scenario, the cellular and D2D SINR are,

respectively,

γ(t)CELLULAR

k,c,n =

∣∣∣h

(t)k,c,n

∣∣∣

2

p(t)k,c,n

C∑

c′ 6=c

K∑

k′

∣∣∣h

(t)k,c′,n

∣∣∣

2

p(t)k′,c′,n

︸ ︷︷ ︸

Interference from cellular links

+C∑

c′

M∑

m′

∣∣∣h

(t)k,tx(m′),c′,n

∣∣∣

2

p(t)rx(m′),tx(m′),c′,n

︸ ︷︷ ︸

Interference from D2D links

+η2

, (2.2)

and

γ(t)D2Drx,c,n =

∣∣∣h

(t)rx(m),tx(m),c,n

∣∣∣

2

p(t)rx(m),tx(m),c,n

C∑

c′

K∑

k′

∣∣∣h

(t)rx(m),c′,n

∣∣∣

2

p(t)k′,c′,n

︸ ︷︷ ︸

Interference from cellular links

+

C∑

c′ 6=c

M∑

m′

∣∣∣h

(t)rx(m),tx(m′),c′,n

∣∣∣

2

p(t)rx(m′),t(m′),c′,n

︸ ︷︷ ︸

Interference from D2D links

+η2

, (2.3)

where:

◮ k is the receiver in a cellular communication;

◮ c is the transmitter in a cellular communication;1The code of the SCM simulator of [45] was developed in the Wireless World Initiative New Radio (WINNER) project.

The software is licensed under the GNU General Public License (GPL).

2.7. Link-to-System Interface 16

◮ n is the PRB;

◮ t is the TTI;

◮ h(t)k,c,n is the channel that models the link between the receiver k and the transmitter c in

PRB n at TTI t;

◮ p(t)k,c,n is the transmit power allocated to transmitter c to link between the receiver k in

PRB n at TTI t;

◮ tx(m) is the transmitter D2D pair m ∈ {0, 1, . . . , R};

◮ rx(m) is the receiver D2D pair m ∈ {0, 1, . . . , R};

◮ p(t)rx(m),tx(m),c,n is the transmit power allocated to transmitter pair tx(m) to link between

the receiver rx(m) in PRB n at cell site c and TTI t;

◮ η2 is the thermal noise power at the receiver.

2.7 Link-to-System Interface

In the following, the link-to-system interface is addressed, which is used to map the

system-level metrics, such as SINR, into link-level performance figures, such as BLock Error

Rate (BLER). The link adaptation selects a proper Modulation and Coding Scheme (MCS) for

each link in order to maximize the throughput for each transmission based on effective gains

achieved by the RRA algorithm [61]. For the sake of simplicity, the MCS for each PRB of a UE

are adapted independently.

Aligned with LTE, a set of fifteen MCSs based on Quadrature Amplitude Modulation (QAM)

and different code rates are available for link adaptation [62]. Figure 2.5 shows the average

throughput curves available for link adaptation, from MCS-1 (leftmost) to MCS-15 (rightmost).

−10 −5 0 5 10 15 20 250

100

200

300

400

500

600

700

800

900

1000

Ave

rage

thro

ugh

pu

t(b

it/s)

SINR (dB)

MCS-1

MCS-2

MCS-3

MCS-4

MCS-5

MCS-6

MCS-7

MCS-8

MCS-9

MCS-10

MCS-11

MCS-12

MCS-13

MCS-14

MCS-15

Figure 2.5: Curves of link-level used for link adaptation.

In each transmission, the link adaptation is determined such that the MCS that yields the

maximum average throughput is selected. SINR thresholds can be found for each MCS, i.e.,

minimal SINR values required to use each MCS. The MCSs considered in this master’s thesis

and its respective SINR thresholds are summarized in Table 2.2.

It should be noted that the link adaptation can be affected by random variations on the

interference levels in the system. We consider that rates are computed considering ideal link

2.8. Imperfect Channel State Information 17

Table 2.2: SINR thresholds for link adaptation [62].

MCS Modulation Code rate [×1024] Rate [Bits/symbol] SINR threshold [dB]

MCS-1 QPSK 78 0.1523 −6.2

MCS-2 QPSK 120 0.2344 −5.6

MCS-3 QPSK 193 0.3770 −3.5

MCS-4 QPSK 308 0.6016 −1.5

MCS-5 QPSK 449 0.8770 0.5

MCS-6 QPSK 602 1.1758 2.5

MCS-7 16-QAM 378 1.4766 4.6

MCS-8 16-QAM 490 1.9141 6.4

MCS-9 16-QAM 616 2.4062 8.3

MCS-10 64-QAM 466 2.7305 10.4

MCS-11 64-QAM 567 3.3223 12.2

MCS-12 64-QAM 666 3.9023 14.1

MCS-13 64-QAM 772 4.5234 15.9

MCS-14 64-QAM 873 5.1152 17.7

MCS-15 64-QAM 948 5.5547 19.7

adaptation following the link level results from Figure 2.5 and that the communications occur

error-free, i.e., there is no packet reception errors and all transmitted data are successfully

received.

2.8 Imperfect Channel State Information

In this master’s thesis, the imperfect CSI issue is addressed in order to illustrate conditions

closer to real-world implementations. The CSI is reported to the transmitter via feedback

channel in which delays can occur, this delay has a negative impact in the system, because the

CSI values are outdated. For example, PC schemes are responsible for mitigating interference

based on channel gain or SINR and when these values do not show the current situation of

the scenario, the system is harmed due to too high or too low transmit power usage.

For the sake of simplicity, it is assumed that all UEs in the system experience the same

delay, which is denoted by an integer number ∆τ of TTIs. Finally, the outdated CSI is given

in ∆τ TTIs, i.e. h(t)k,c,n = h

(t+∆τ)k,c,n . This is the CSI effectively used as CSI. Figure 2.6 shows two

cases, the first case has ∆τ = 0 (perfect CSI), and the second case has ∆τ = 2 (imperfect CSI).

In the first TTI of the simulations all values are available to computer CSI in both case.

In the first case, when ∆τ = 0 the PC schemes determine transmit power based on current

TTI, because it does not have delay. In the second case, when ∆τ = 2 the PC schemes

determine transmit power based on the CSI of a past TTI. For example, the CSI of the 2nd

and 3th TTI are based on the CSI of the 1st TTI, while the CSI of the 5th and 6th TTI are based

on the CSI of the 4th TTI, and so on.

1 432 5 876 9 10

Fee

dbac

k

Beg

in

Fee

dbac

k

Fee

dbac

k

Fee

dbac

k

Fee

dbac

k

Fee

dbac

k

Fee

dbac

k

Fee

dbac

k

Fee

dbac

k

1 432 5 876 9 10

No

Fee

dba

ck

Beg

in

Fee

dbac

k

Fee

dbac

k

Fee

dbac

k

No

Fee

dbac

k

No

Fee

dbac

k

No

Fee

dbac

k

No

Fee

dbac

k

No

Fee

dbac

k

No delay

Delay

Figure 2.6: Imperfect CSI using feedback delay.

2.9. System Level Simulation 18

2.9 System Level Simulation

Computer simulation is taken as an important tool to analyze and assess the performance

of complex systems such as D2D links. Thus, a system-level simulation tool based on

the system model described in this chapter has been implemented. The main parameters

considered in the simulations are summarized in Table 2.3.

Table 2.3: Simulation parameters for urban-macrocell and microcell environments.

Parameters Urban-macrocell Urban-microcell Unit

Cellular scenario

Number of eNBs 7 7 -Inter-site distance 3 000 500 meNB height 32 12.5 mUE height 1.5 1.5 mCSI knowledge Perfect Perfect / Imperfect -Link adaptation LTE (15 MCSs) LTE (15 MCSs) [63]Interference margin Last interference Last interference [3]eNB transmit power 48 38 dBm [64]UE transmit power 24 24 dBm [64]Thermal noise power −112.4 −112.4 dBm [64]SINR threshold for lowest MCS −6.2 −6.2 dB [65]Average user speed 3 3 km/h [64]

OFDMA

Central carrier frequency 1.9 1.9 GHz [66]System bandwidth 5 5 MHzNumber of PRB 25 25Number of symbols per TTI 14 14

Propagation

Path loss model for cellular links 34.5 + 35 log10(d) 35.7 + 38 log10(d) dBPath loss model for D2D links 37 + 30 log10(d) 37 + 30 log10(d) dBLognormal shadowing std. deviation 8 10 dBFast fading model 3GPP SCM 3GPP SCM [66]

Simulation

Traffic model Full buffer Full bufferNumber of UEs per cell 4 4,8,16,32Snapshot duration 1 s 1 s

2.10 Classification of Metrics Used in Energy Efficiency

We need to understand some metrics used in this master’s thesis, which allow for the

comparison of different algorithms. According to [2], energy efficiency metrics are used to

describe the ability of a telecommunication system to minimize energy waste. For instance,

when a telecommunication system transmits more data (bits) with less power (Watt), this

system is considered more energy efficient. An energy efficiency metric can be defined at the

network level, the system level and the component level. In Table 2.4 are summarized the

main metrics to energy efficiency.

2.10.1 Energy Efficiency at the Network Level

Network level metrics are used to evaluate the energy efficiency of an entire network

or part of it. Network level metrics assess energy efficiency at the network level by

considering the features and properties of the capacity and coverage of the network. In other

words, it is normally used to evaluate a network for internal operator use or to satisfy an

environmental assessment. So, the network level is considered a metric that will cover not only

one equipment, but also a telecommunication network composed of different interworking

equipments.

2.10. Classification of Metrics Used in Energy Efficiency 19

2.10.2 Energy Efficiency at the System Level

System level metrics are related with access node, which is used to compare and analyze

RRM algorithms that approach resource allocation, power control, interference coordination

and cooperative scheduling. Important system level metrics used in this master’s thesis are

summarized bellow:

◮ Spectral efficiency is a metric that considers the amount of information tha can be

transmitted per bandwidth unit. Spectral efficiency is directly related with system

capacity, therefore, it is possible to compare capacity of two or more algorithms in a

system providing the same service by using it. In this master’s thesis spectral efficiency

is expressed in [bps/Hz/cell];

◮ Power efficiency is a essential metric, which can be modified according to RRM

algorithms, load, and environmental factors. Power efficiency can be describe as spectral

efficiency per unit of power, i.e., [bps/Hz/cell/W].

2.10.3 Energy Efficiency at the Component Level

Component level metrics can be used in the design, development and manufacture of

energy efficient devices. Component level metrics are useful to compare the hardware

of a communication device, such as Central Processing Unit (CPU), memory, power

source, Large-Scale Integration (LSI) microfabrication and power amplifier. Measuring and

understanding the energy efficiency or energy consumption of each component within the

equipment helps to identify key components in a system with regard to energy saving.

Table 2.4: Metrics Used in Energy Efficiency.

Level Units Description

Component Level

W/Gbps The ratio of energy consumption to effective system capacityGbps/W The ratio of useful work to power consumptionMIPS/W Millions of instructions per second per Watt

MFLOPS/W Millions of floating-point operations per second per watt

System Level

b/s/Hz Rate of information can be transmitted in a bandwidthb/s/Hz/W The spectral efficiency per Watt

(b·m)/s/Hz/W Rate of transmission and the transmission distance attainablefor a given bandwidth and power resources supplied

J/bit Number of bits transmitted per Joule of energy

Network Level

km2/W The ratio of coverage area to site power consumptionW/km2 The power consumed per unit area

Users/W The ratio of users served during the peak traffic hourJ/bit/m2 Number of transferred bits and the coverage area

W/bps/m2 The average power usage with respect to the averagetransmission rate and the coverage area

Chapter 3Energy Efficiency RRM Methods

This chapter covers the Power Control (PC) schemes and downtilt used in this master’s

thesis, so that a reader can understand the principle of Equal Power Allocation (EPA), Fixed

Power, Open Loop Power Control (OLPC), Closed Loop Power Control (CLPC), Soft Dropping

Power Control (SDPC), Closed Loop Soft Dropping (CLSD) and downtit. The remainder of this

chapter is structured as follows. In Section 3.1.1 is described the baselines EPA and Fixed

Power. Next, in the Sections 3.1.2, 3.1.3 and 3.1.4 are detailed LTE PC schemes, SDPC and

CLSD, respectively. Finally, in Section 3.2 downtilt is described.

3.1 PC

In this section, the PC schemes applied in this master’s thesis are described. The baseline

algorithms are the EPA and the Fixed Power. These algorithms have important function in

this master’s thesis, because they are the reference to the other PC schemes studied in this

master’s thesis.

3.1.1 EPA and Fixed Power

EPA is characterized by the equal distribution of the total transmit power PeNB or PUE

among the Physical Resource Block (PRB). In other words, the EPA scheme obtains the power

pk,n for each User Equipment (UE) k at PRB n as

Pk,n =

PeNB/NPRB, for eNB transmitters,

PUE/Nk, for UE transmitters.(3.1)

where Nk is the number of PRBs scheduled to the UE. Fixed Power is a simple PC scheme,

where pk,n = PUE/NPRB = 10 dBm, ∀k, n.

3.1.2 LTE Power Control

The OLPC and CLPC are the standard LTE power control algorithms which work with

fractional path loss compensation [63]. For this algorithm, the total transmit power pk of a

cellular or D2D UE k is given as

pk = min{PUE, P0 − αG+ 10 log10 Nk+ △}, (3.2)

where PUE is the maximum UE power, 0 ≤ α ≤ 1 is the pathloss compensation factor, G

denotes the path gain of the channel, Nk is the number of PRBs scheduled to UE k and △ is a

dynamic offset. This dynamic offset differentiates OLPC from CLPC, because OLPC does not

3.1. PC 21

have feedback and, therefore, △= 0, while CLPC has a feedback which can be computed as

△=

(Γk − γk)σ, if (Γk − γk)σ > 1,

1, if (Γk − γk)σ < 1.(3.3)

where 0 < σ ≤ 1 is the dynamic offset compensation factor. P0 is power level used to control

the target SNR Γk, which is given according to [67] as

P0 = α(Γk + PN) + (1− α)(PUE − 10 log10 Nk), (3.4)

where, for simplicity, PN is the thermal noise power at the cellular or D2D receiver,

respectively, eNB or UE. After total transmit power is updated, the power pk,n in each PRB n

according to the EPA scheme as

pk,n = pk/Nk. (3.5)

3.1.3 Soft Dropping Power Control (SDPC)

Power Control (PC) with variable SINR is an alternative approach to protect cellular and

D2D communications from mutual interference. This approach, in which the target Signal

to Interference-plus-Noise Ratio (SINR) gradually decreases as the required transmit power

rises, it is called Soft Dropping (SD) in [42,68–70]. It increases the probability of configuring a

feasible PC problem — in which the target SINR values of all co-channel links can be reached

— since links with worse quality, which demand higher power, aim at lower SINR values while

links with better quality, which demand lower power, aim at higher SINR values.

The principle of the SD algorithm is illustrated in the Figure 3.1.

p [dBm]

Γ [dB]

Pmin

Γmax

Pmax

Γmin

pk,n

Γk,n(pk,n)

Figure 3.1: Target SINR as function of a variable transmit power.

In the SDPC scheme, the transmit power per PRB of each link is iteratively adjusted in

order to find a power vector p for all UEs in the system such that the SINR γk,n of each UE k

in PRB n satisfies

γk,n(p) ≥ Γk,n(pk,n), (3.6)

where Γk,n(pk,n) is the target SINR of the UE k in the PRB n, which varies according to the

required transmit power pk,n.

The SDPC scheme uses a target SINR varying from a maximum value Γmax to a minimum

Γmin as the required transmit power goes from a minimum value Pmin to a maximum Pmax.

Here, the range ∆P = Pmax − Pmin is termed the PC range. For pk,n ≤ Pmin, one attempts to

maintain a high quality connection by aiming at a target SINR Γmax. For pk,n ≥ Pmax, one

3.1. PC 22

aims at a target SINR Γmin which is relatively easier to reach when channel conditions are

bad. Finally, for Pmin < pk,n < Pmax, one aims for a target SINR Γk,n(pk,n) that linearly (in

logarithmic scale) trades SINR for transmit power. The target SINR Γk,n(p(t)k,n) of UE k in the

PRB n at Transmission Time Interval (TTI) t is given according to

Γk,n(p(t)k,n) =

Γmax, p(t)k,n ≤ Pmin,

Γmax

(

p(t)k,n

Pmin

, Pmin < p(t)k,n < Pmax,

Γmin, p(t)k,n ≥ Pmax,

(3.7)

where

ρ =log10(Γmin/Γmax)

log10(Pmax/Pmin). (3.8)

Then, the power per PRB of each UE is updated every transmission as follows

p(t+1)k,n = p

(t)k,n

(

Γk,n(p(t)k,n)

γk,n(p(t))

, (3.9)

where β is a control parameter given by (1 − ρ)−1 [68].

Finally, whenever the achieved power p(t+1)k,n is over Pmax or under Pmin, it is constrained as

follows

p(t+1)k,n = min{Pmax,max{p

(t+1)k,n , Pmin}}. (3.10)

In this master’s thesis, the maximum power Pmax is exactly the power that would be

obtained in each resource by employing EPA among the total number of resources as

follows [43]:

Pmax =

PeNB/NPRB, for eNB transmitters,

PUE/NPRB, for UE transmitters.(3.11)

3.1.4 Closed Loop Soft Dropping (CLSD)

The CLSD is a hybrid PC scheme, because it uses features of CLPC and SDPC. The total

transmit power pk,n of each cellular or D2D UE k in PRB n at TTI t is given according to

p(t+1)k,n = min{PUE, P0 − αG+ 10 log10 Nk+ △}, (3.12)

where PUE is the maximum UE power, 0 ≤ α ≤ 1 is the pathloss compensation factor, G

denotes the path gain of the channel, Nk is the number of PRBs scheduled to the UE k, and

P0 is power level used to control the target SINR Γk,n(p(t)k,n), which is given as

P0 = α(Γk,n(p(t)k,n) + PN) + (1− α)(PUE − 10 log10 Nk), (3.13)

where, for simplicity, PN is the thermal noise power at the cellular or D2D receiver,

respectively, Evolved Node B (eNB) or UE and △ is a dynamic offset, which can be written

as

△=

(

Γk,n(p(t)k,n)− γk,n(p

(t)))

β, if(

Γk,n(p(t)k,n)− γk,n(p

(t)))

β > 1,

1, if(

Γk,n(p(t)k,n)− γk,n(p

(t)))

β < 1.(3.14)

where 0 ≤ β ≤ 1 is the dynamic offset compensation factor. The target SINR Γk,n(p(t)k,n) of UE k

in the PRB n at TTI t and ρ are given according Equations (3.7) and (3.8), respectively.

3.2. Downtilt 23

In this master’s thesis, the maximum power Pmax is given according

Pmax = PUE/NPRB (3.15)

3.2 Downtilt

Currently, wireless systems face several issues that must be considered in its development

and optimization. We can mention co-channel interference, irregular geographical terrain and

improper antenna position as some factors that can have a negative impact on the network

performance.

One simple and efficient method to reduce some of these negative effects is called downtilt.

This method is used to adjust the coverage radius of an eNB and reduce co-channel

interference by increasing cell isolation. There exist many different downtilt schemes, for

example, mechanical tilt, electrical tilt, Variable Electrical Tilt (VET) and Remote Electrical

Tilt (RET), which can be used to adjust coverage area, cell load, improve system capacity and

traffic distribution.

In [71], the authors show important concepts about downtilting and the relationship

between antenna height, downtilt angle, and coverage radius. The study outcomes that due

to the severe urban propagation environments, the coverage area control by antenna downtilt

has been reduced due to the high rise of tall buildings.

In [72], the authors discuss the impact of the Base Station (BS) mechanical antenna

downtilt scheme on the downlink capacity of a 6-sectored Wideband Code Division Multiple

Access (WCDMA) cellular network considering a macro-cellular environment. They conclude

that an optimum mechanical downtilt angle exists in all simulation scenarios, and clearly this

angle can be defined for each site and antenna configuration separately, depending on the BS

antenna height and vertical beamwidth together with the site spacing. In relation to capacity,

the downlink capacity increases with the downtilt angle but the coverage is reduced.

3.2.1 Antenna Fundamentals

Antenna is a device used for converting electromagnetic radiation in space into electrical

currents in conductors or vice-versa, depending on whether it is being used for reception or

for transmission, respectively. The pattern in which the radiating wave travels in the free

space can be controlled by using different antenna parameters. The main parameters used

in this master’s thesis are antenna azimuth orientation and antenna downtilt. These antenna

parameters that define the radiation pattern are explained below:

◮ Antenna azimuth orientation is the direction of the main lobe in the horizontal

direction with positive values for the clockwise measurements from the horizontal axis.

◮ Antenna downtilt is the direction of the main lobe in the vertical direction with positive

values for the down side tilting of the main lobe. The downtilt of the antenna radiation

pattern can be done either by mechanical downtilt or by electrical downtilt. In case of

mechanical downtilt, with changes in the downtilt values, there will be a variation in the

horizontal radiation pattern of the antenna. In electrical downtilt, only vertical antenna

radiation pattern is affected. In this master’s thesis, the terms downtilt and electrical

downtilt are used interchangeably.

This concept can be easier understood in Figure 3.2, which shows a macrocell scenario

with eNB in center and 6 UEs. Each cell site have an azimuth orientation, which are 60°,

3.2. Downtilt 24

180°and 300°. Another parameter in this scenario is the antenna downtilt, with each cell site

having 20°downtilt angle with respect to the horizontal direction.

������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������

20°

60º

Azimuth Orientation

Downtilt

Figure 3.2: Azimuth orientation and downtilt in a macrocell scenario.

3.2.2 Electrical Antenna Downtilt

In this master’s thesis, electrical downtilt is used, which has some differences compared

to mechanical downtilt. The mechanical downtilt uses specific accessories, which are

responsible for modifying the antenna bracket, while electrical downtilt changes the phase

of the input signal and consequently the signal propagation directions. Each technique has

clear differences in antenna radiation. The coverage is reduced in central direction with

mechanical downtilt. However, the coverage in side directions is increased. When electrical

downtilt is used, the coverage has an uniform reduction in the direction of the antenna

azimuth orientation and the gain is reduced uniformly.

Antenna models were created to analyze electrical and mechanical downtilt in cellular

networks. In [66], only horizontal radiation patterns were used. Nevertheless, several

papers have described the improvement that can be achieved with the addition of the vertical

pattern [44, 46, 47, 49]. In this master’s thesis, a simple model for the vertical antenna

pattern proposed in [73] is used, which is an extension of the 3rd Generation Partnership

Project (3GPP) model. The horizontal model of antenna pattern in 3GPP [66] has a maximum

gain Gm = 14 dB, front to back ratio FRBh = 20 dB and a horizontal half power beamwidth

HPBWh = 70°. The horizontal antenna gain equation can be written as

Gh(ϕ) = −min{12(ϕ/HPBWh)2, FRBh}+Gm (3.16)

where ϕ, −180° ≤ ϕ ≤ 180°, is the azimuth in degrees. It is possible to see that the model does

not have antenna tilt, since it requires an antenna radiation pattern model defined over both

horizontal and vertical directions. In [73], not only other parameter values for Gm = 18 dB,

FRBh = 30 dB and HPBWh = 65° are selected, but also the vertical pattern is defined as

Gv(θ) = max{−12((θ− θtilt)/HPBWv)2, SSLv} (3.17)

where θ, −90° ≤ θ ≤ 90°, is the angle relative to the horizontal plane. The others parameters

are the electrical downtilt angle θtilt, side lobe level SSLv = 18 dB and vertical half power

beamwidth HPBWv = 6.2°. These parameters are defined based on the Kathrein 742215 data

3.2. Downtilt 25

sheet described in [73]. Through the combination of horizontal and vertical gain, it is possible

to get the antenna gain in a general direction (ϕ, θ) as

G(ϕ, θ) = Gh(ϕ) +Gv(θ). (3.18)

Chapter 4Results and Analysis

This chapter covers the results of Power Control (PC) schemes and antenna downtilt used

in this master’s thesis for different scenarios. The remainder of this chapter is structured

as follows. In Section 4.1.1 direction are presented and discussed the results to micro-cell

scenario for Downlink (DL). In the Section 4.1.2 is detailed micro-cell scenario for Uplink (UL).

Finally, in Section 4.2 the effect of antenna downtilt in a macro-cell scenario is discussed.

4.1 Power Control

To understand the behavior of PC schemes in a cellular network with underlaying

Device-to-Device (D2D) communications, it is important to analyze both DL and UL scenarios,

select PC schemes to cellular and D2D communications and adjust PC parameters based on

spectral efficiency and power efficiency.

There are several parameters in PC schemes that need to be verified, modified and updated

depending on the scenarios of interest. This section will present several scenarios, such

as macro-cell or micro-cell scenarios with different number of users and using perfect or

imperfect Channel State Information (CSI).

4.1.1 Power Control Evaluation in a Micro-cell Scenario (Downlink)

This section provides the performance assessment of a PC algorithm with variable Signal

to Interference-plus-Noise Ratio (SINR) for cellular and D2D communications in a multi-cell

scenario using DL direction. Results are obtained through system-level simulations aligned

with 3rd Generation Partnership Project (3GPP) Long Term Evolution (LTE) architecture [60,

63,64,66]. The main parameters considered in the simulations are summarized in Table 2.3,

a load of 4 users are considered.

It is important to calibrate the Soft Dropping Power Control (SDPC), otherwise, the SDPC

can harm the system. The results are compared with the baseline Equal Power Allocation

(EPA).

In the following, while the maximum target SINR is fixed in Γmax = 25 dB, which is higher

than the SINR threshold of the highest MCS [65], and the maximum power per Physical

Resource Block (PRB) Pmax is calculated by Equation (3.11), the PC range ∆P and the

minimum target SINR Γmin are varied for calibration purposes. The simulated minimum

SINR values Γmin are above the SINR threshold of the lowest MCS [65] while the simulated

minimum output power values Pmin of a User Equipment (UE) do not go below −40 dBm [74].

Figure 4.1 shows the total system spectral efficiency and the average transmit power achieved

by the Soft Dropping (SD) algorithm for cellular and D2D communications.

4.1. Power Control 27

010

2030

40

−50

510

15202.8

3

3.2

3.4

3.6

3.8

PC range [dB]Min. target SINR [dB]

To

tal

syst

em s

pec

tral

eff

icie

ncy

[b

ps/

Hz/

cell

]

P1

(a) Total system spectral efficiency achieved by cellularand D2D receivers (SD in cellular transmitters andEPA in D2D transmitters).

010

2030

40

−50

510

15203.2

3.4

3.6

3.8

4

PC range [dB]Min. target SINR [dB]

To

tal

syst

em s

pec

tral

eff

icie

ncy

[b

ps/

Hz/

cell

]

P2

(b) Total system spectral efficiency achieved by cellularand D2D receivers (SD in D2D transmitters and EPAin cellular transmitters).

010

2030

40

−50

510

152034

35

36

37

38

PC range [dB]Min. target SINR [dB]

Tra

nsm

it p

ow

er [

dB

m]

(c) Transmit power consumed by cellular transmitters(SD in cellular transmitters and EPA in D2Dtransmitters).

010

2030

40

−50

510

152014

16

18

20

22

24

PC range [dB]Min. target SINR [dB]

Tra

nsm

it p

ow

er [

dB

m]

(d) Transmit power consumed by D2D transmitters(SD in D2D transmitters and EPA in cellulartransmitters).

Figure 4.1: Calibration of SD algorithm regarding the total system spectral efficiency and transmitpower.

As depicted in Figures 4.1(a) and 4.1(b), the highest total system spectral efficiency values

are achieved for Γmin = 20 dB while the lowest transmit power values are achieved for Γmin =

−5 dB in both cellular and D2D communication cases. Therefore, I consider two operation

points: P1 is set for evaluating system spectral efficiency gains, while P2 is set for evaluating

the power saving.

For Γmin = 20 dB, the gains in system spectral efficiency practically saturate for ∆P greater

than 30 dB, as also shown in Figures 4.1(a) and 4.1(b). Thus, P1 is set as (Γmin = 20 dB,

∆P = 30 dB).

Considering 5% of reduction on the total system spectral efficiency achieved when applying

SD to D2D communications, P2 is set as (Γmin = −5 dB, ∆P = 30 dB), hereafter P1 and P2 will

be used for the remaining results.

The relative gains in terms of total system spectral efficiency and power saving achieved

by applying the SD algorithm to cellular and D2D communications in comparison to the

EPA scheme are summarized in Table 4.1 for the two considered operation points. The

application of SD in D2D communications always provides a better relative performance for

both operation points. As expected, while the operation point P1 provides better relative gains

for the total system spectral efficiency, P2 performs better in power saving.

In order to protect the cellular communications, the SD algorithm applied to cellular

4.1. Power Control 28

Table 4.1: Relative gains of performance by applying the SD algorithm to cellular and D2Dcommunications.

Total spectral eff. gain Power saving

P1 P2 P1 P2

SD in cellular transmitters +1% −19% 5% 57%SD in D2D transmitters +7% − 5% 49% 84%

transmitters is set to use the operation point P1 and the SD algorithm applied to D2D

transmitters is set to use P2. Figure 4.2 presents the Cumulative Distribution Function (CDF)

of SINR and interference power perceived by cellular and D2D receivers.

−30 −20 −10 0 10 20 30 40 50 600

20

40

60

80

100

SINR [dB]

CD

F [

%]

EPA (cellular) x EPA (D2D)

SD (cellular) x EPA (D2D)

EPA (cellular) x SD (D2D)

SD (cellular) x SD (D2D)

(a) SINR of cellular communications.

−30 −20 −10 0 10 20 30 40 50 600

20

40

60

80

100

SINR [dB]C

DF

[%

]

EPA (cellular) x EPA (D2D)

SD (cellular) x EPA (D2D)

EPA (cellular) x SD (D2D)

SD (cellular) x SD (D2D)

(b) SINR of D2D communications.

−110 −100 −90 −80 −700

20

40

60

80

100

Interference power [dBm]

CD

F [

%]

EPA (cellular) x EPA (D2D)

SD (cellular) x EPA (D2D)

EPA (cellular) x SD (D2D)

SD (cellular) x SD (D2D)

(c) Interference power of cellularcommunications.

−110 −100 −90 −80 −700

20

40

60

80

100

Interference power [dBm]

CD

F [

%]

EPA (cellular) x EPA (D2D)

SD (cellular) x EPA (D2D)

EPA (cellular) x SD (D2D)

SD (cellular) x SD (D2D)

(d) Interference power of D2D communications.

Figure 4.2: SINR and interference power of cellular and D2D communications by applying SD and EPAschemes.

Observing Figure 4.2(a), when the SD algorithm is applied to cellular communications the

SINR and interference curves are practically maintained in comparison to those obtained

using EPA. Following results shown in Table 4.1, the operation point P1 has the best

performance in terms of total system spectral efficiency. For this point, the highest SINR levels

achieved by cellular communications are marginally reduced while D2D communications

maintain the same SINR levels, as shown in Figure 4.2(b). In addition, the power reduction of

cellular transmitters does not contribute to the reduction of the interference power perceived

by both cellular and D2D receivers, as shown in Figures 4.2(c) and 4.2(d). In general,

interfering cellular transmitters are far away from D2D receivers (which only happen to be

inside hotspot zones at cell-edges) and from cellular receivers located in other cells.

When the SD algorithm is applied to D2D communications the high SINR levels achieved

by D2D communications are reduced, as shown in Figure 4.2(b), and the SINR levels of

cellular communications are considerably improved, as presented in Figure 4.2(a). This

4.1. Power Control 29

occurs because, in general, D2D transmitters act as interfering sources quite close to cellular

receivers while cellular transmitters are quite distant from D2D receivers since these are

inside hotspot zones at cell-edges. Besides that, D2D receivers are more distant from their

interfering D2D transmitters, which are regularly distributed over the multi-cell coverage area

at cell-edges, than cellular receivers.

When the SD algorithm is applied to both cellular and D2D communications at once, it is

possible to notice only tiny gains on the reduction of interference power levels, as shown in

Figures 4.2(c) and 4.2(d).

Figure 4.3 presents the system spectral efficiency for cellular communications, D2D

communications and both communications modes considering the SD algorithm applied to

D2D communications in both operation points P1 and P2.

Cellular D2D Total0

0.5

1

1.5

2

2.5

3

3.5

4

Communications

Sy

stem

sp

ectr

al e

ffic

ien

cy [

bp

s/H

z/ce

ll]

EPA (cellular) x EPA (D2D)

EPA (cellular) x SD (D2D) − P1

EPA (cellular) x SD (D2D) − P2

Figure 4.3: System spectral efficiency by applying SD and/or EPA to cellular algorithms and/or D2Dtransmitters.

As observed in Figure 4.3, the cellular communications always have their performance

improved when the SD algorithm is applied to D2D communications for both operation

points P1 and P2. For the operation point P1, there is a reduction of 49% on the power of

D2D transmitters, as shown in Table 4.1, while the performance of D2D communications

is practically maintained. The system spectral efficiency relative gains by applying the SD

algorithm to D2D communications in comparison to the EPA scheme are summarized in

Table 4.2.

Table 4.2: Relative gains of system spectral efficiency by applying the SD algorithm to D2Dcommunications.

Cellular gain D2D gain Total gain

P1 +14% 0% +7%P2 +39% −44% −5%

As shown in Table 4.2, there is a considerable improvement on the system spectral

efficiency of cellular communications for both operation points, which is accompanied with

high reduction on the transmit power of D2D transmitters (49% for P1, as mentioned before,

and 84% for P2) as shown in Table 4.1. The main reason for the gains is due to the SDPC has

a range of target SINR, while EPA provides a fixed transmit power.

4.1. Power Control 30

4.1.2 Power Control Evaluation in a Micro-cell Scenario (Uplink)

Section 4.1.1 showed results regarding the DL. The focus of this section is the UL in a

Micro-cell scenario, which is aligned with the LTE architecture [60, 63, 64, 66]. The main

parameters considered in the simulations are summarized in Table 2.3, a load of 4 users are

considered.

4.1.2.1 LTE PC schemes and SDPC

020

4060

−50

510

15202.4

2.6

2.8

3

3.2

3.4

3.6

PC range [dB]Min. SINR [dB]

To

tal

spec

tral

eff

. [b

ps/

Hz/

cell

]

(a) Total spectral efficiency (SDPC in cellular andNo-PC in D2D links).

020

4060

−50

510

1520

5

10

15

20

PC range [dB]Min. SINR [dB]

Po

wer

eff

icie

ncy

[b

ps/

Hz/

cell

/W]

(b) Cellular Power efficiency (SDPC in cellularand No-PC in D2D links).

020

4060

−50

510

15202.4

2.6

2.8

3

3.2

3.4

3.6

PC range [dB]Min. SINR [dB]

Tota

l sp

ectr

al e

ff.

[bps/

Hz/

cell

]

(c) Total spectral efficiency (No-PC in cellularand SDPC in D2D links).

020

4060

−50

510

1520

5

10

15

PC range [dB]Min. SINR [dB]

Po

wer

eff

icie

ncy

[b

ps/

Hz/

cell

/W]

(d) D2D Power efficiency (No-PC in cellular andSDPC in D2D links).

Figure 4.4: Calibration of the SDPC scheme by applying it to cellular or D2D links. The PC range∆P = 0dB gives the performance of fixed power approach. Minimum target SINR values aresimulated until Γmin = −5 dB because the SINR threshold of the lowest MCS is −6.2 dB.

In this section, SDPC, Open Loop Power Control (OLPC), Closed Loop Power Control (CLPC)

are step by step analyzed. At first, the SDPC and the OLPC are calibrated and evaluated.

After CLPC is calibrated based on the results obtained from OLPC. Finally, all PC schemes

are evaluated and compared to EPA and Fixed Power. For performance evaluation, the

energy efficiency is measured using the power efficiency metric [2], which gives the ratio

of the total system spectral efficiency achieved by cellular and D2D communications to the

average transmit power in [bps/Hz/cell/W]. As baseline, the No-PC approach with EPA among

scheduled PRBs, i.e., pk,n = PUE/Nk, and the fixed power approach with pk,n = PUE/NPRB =

10 dBm, ∀k, n are considered.

For calibration purposes, Figures 4.4 and 4.5 show the performance of the SDPC and

OLPC, respectively. For each PC scheme, there are operating points responsible for high

energy efficiency and reasonable total system spectral efficiency gains.

In order to protect cellular communications, operating points are chosen for each PC

scheme that maintain the total spectral efficiency or achieve the highest power efficiency

gains for D2D communications. The relative performance gains and the considered operating

4.1. Power Control 31

00.2

0.40.6

0.81

10

15

20

252.4

2.6

2.8

3

3.2

3.4

3.6

αTarget SNR [dB]

To

tal

spec

tral

eff

. [b

ps/

Hz/

cell

]

(a) Total spectral efficiency (OLPC in cellularand No-PC in D2D links).

00.2

0.40.6

0.81

10

15

20

255

10

15

αTarget SNR [dB]

Pow

er e

ffic

iency

[bps/

Hz/

cell

/W]

(b) Cellular Power efficiency (OLPC in cellularand No-PC in D2D links).

00.2

0.40.6

0.81

10

15

20

252.4

2.6

2.8

3

3.2

3.4

3.6

αTarget SNR [dB]

To

tal

spec

tral

eff

. [b

ps/

Hz/

cell

]

(c) Total spectral efficiency (No-PC in cellularand OLPC in D2D links).

00.2

0.40.6

0.81

10

15

20

250

10

20

30

αTarget SNR [dB]P

ow

er e

ffic

iency

[bps/

Hz/

cell

/W]

(d) D2D Power efficiency (No-PC in cellular andOLPC in D2D links).

Figure 4.5: Calibration of the OLPC scheme by applying it to cellular or D2D links. The pathlosscompensation factor α = 0 gives the No-PC performance.

points for each PC scheme are summarized in Table 4.3. While the SDPC scheme performs

better for cellular communications, the OLPC scheme is better for D2D communications in

terms of power efficiency. Even when OLPC uses, e.g., an operating point (Γk = 10 dB, α = 0.3)

(not shown in the table) providing a total system spectral efficiency loss of 9% like the SDPC,

its power efficiency gain of 297% is even higher than the 114% of the SDPC scheme.

Table 4.3: Relative gains by applying SDPC and OLPC to cellular or D2D links in comparison to theNo-PC approach (%).

Total spectral efficiency loss Power efficiency gain

SDPC OLPC SDPC OLPC

Cellular links 0 0 79 3D2D links 9 20 114 379

Operating points for cellular communications:SDPC (Γmin = 20 dB, ∆P = 10 dB), OLPC (Γk = 25 dB, α = 0.3)

Operating points for D2D communications:SDPC (Γmin = −5 dB, ∆P = 20 dB), OLPC (Γk = 10 dB, α = 0.5)

In order to understand the power efficiency gains presented in Table 4.3, Figure 4.6 shows

the CDFs of the SINR and interference power levels obtained by applying SDPC and OLPC

schemes to cellular or D2D links. By comparing the SDPC scheme to fixed power approach

both applied to cellular links, the SDPC scheme only improves the highest SINR levels of

cellular links, as shown in Figure 4.6(a). As the SDPC scheme uses the maximum power for all

UEs with SINR values below the minimum SINR target, their transmit power values are fixed

by Equation (3.11) as in the fixed power approach. The No-PC approach achieves higher SINR

levels than the fixed power approach, but its interference power levels are also higher. On its

4.1. Power Control 32

turn, the OLPC scheme reduces the SINR levels of D2D links, because the power reduction

does not considerably improve their interference power levels, as shown in Figure 4.6(d), but

it provides the closest performance for cellular links compared with the conventional scenario,

see Figures 4.6(a) and 4.6(c).

−30 −20 −10 0 10 20 30 40 500

20

40

60

80

100

SINR [dB]

CD

F[%

]

EPA (cellular)

Fix Pwr (cellular) x EPA (D2D)

SDPC (cellular) x EPA (D2D)

EPA (cellular) x EPA (D2D)

EPA (cellular) x OLPC (D2D)

(a) SINR of cellular communications.

−30 −20 −10 0 10 20 30 40 500

20

40

60

80

100

SINR [dB]

CD

F[%

]

Fix Pwr (cellular) x EPA (D2D)

SDPC (cellular) x EPA (D2D)

EPA (cellular) x EPA (D2D)

EPA (cellular) x OLPC (D2D)

(b) SINR of D2D communications.

−130 −120 −110 −100 −90 −80 −70 −600

20

40

60

80

100

Interference power [dBm]

CDF[%]

EPA (cellular)

Fix Pwr (cellular) x EPA (D2D)

SDPC (cellular) x EPA (D2D)

EPA (cellular) x EPA (D2D)

EPA (cellular) x OLPC (D2D)

(c) Interference power of cellular communications.

−130 −120 −110 −100 −90 −80 −70 −600

20

40

60

80

100

Interference power [dBm]

CDF[%]

Fix Pwr (cellular) x EPA (D2D)

SDPC (cellular) x EPA (D2D)

EPA (cellular) x EPA (D2D)

EPA (cellular) x OLPC (D2D)

(d) Interference power of D2D communications.

Figure 4.6: SINR and interference power levels by applying SDPC and OLPC schemes to cellular orD2D links. No-PC and fixed power approaches are considered as baselines. No-PC (cellular)represents the conventional scenario without D2D communications underlaying the cellularnetwork.

As the SDPC scheme aims mainly at the reduction of high SINR levels (which reduces the

power consumption without significantly harming the system spectral efficiency) while the

OLPC scheme compensates the pathloss even for low SINR levels, the SDPC scheme provides

a better power efficiency for cellular communications. For D2D communications, the OLPC

scheme achieves a reduced power consumption by exploiting the UEs’ physical proximity. As

the OLPC scheme applied to D2D links provides the highest energy efficiency gains, Figure 4.7

presents the system spectral efficiency of cellular and D2D communications by applying OLPC

to D2D links and no PC for cellular links.

As it can be seen, the factor α can be used to control the performance trade-off between

cellular and D2D communications. We also see that for α = 1.0, which provides the lowest

possible transmit power levels for D2D transmitters (5 dBm), the system spectral efficiency

for D2D communications is practically zero. It means that D2D transmitters are introducing

interference to the system but D2D receivers are not achieving the SINR threshold of the

lowest MCS to attain communication. Thus, the minimum cost for enabling system spectral

efficiency gains for D2D communications considering the most favorable scenario for sharing

resources in all cells represents a minimal impact of 11% on the system spectral efficiency

of cellular UEs. To get gains in the total system spectral efficiency over the conventional

4.1. Power Control 33

0.5

1

1.5

2

2.5

3

3.5

To

tal

syst

em s

pec

tral

eff

icie

ncy

[b

ps/

Hz/

cell

]

Cellular communications

D2D communications

11%

Pathloss compensation factor α0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Conv.

Figure 4.7: Total system spectral efficiency by applying OLPC to D2D links without PC for cellular links(No-PC approach). The pathloss compensation factor α of the OLPC scheme is varied fortarget SNR Γk = 10 dB. The conventional scenario considers the No-PC approach in itscellular links.

scenario, α should be lower than 0.6. For α = 0.5, the impact is 13%, but the power efficiency

is maximum, as it can be seen in Figure 4.5(d). The highest impact (which is 30%) is obtained

for α = 0, i.e., when the maximum power PUE is employed to D2D links as in the No-PC

approach.

The relative performance gains of OLPC for D2D links by varying α are summarized in

Table 4.4. Most α values provide huge power saving gains (measured against the total transmit

power used for transmission) but low α values are preferred to avoid high system spectral

efficiency losses. To achieve high power efficiency gains for D2D links, α should be {0.4, 0.5}.

Table 4.4: Relative performance gains of OLPC for D2D links compared with no-PC for D2D links (%).

α 0.1 0.2 0.3 0.4 0.5 0.6

Spectral efficiency loss 13 28 43 59 73 84Power saving gain 52 75 86 91 94 96Power efficiency gain 81 188 298 373 379 311

Using the same parameters of Table 4.3, it is possible to analyze the CLPC, which is

another LTE PC scheme. The parameter α = 0.5 is defined as the best value to provide power

efficiency at OLPC, see Figure 4.5 and the same α is used at CLPC. The CLPC has a new

parameter called dynamic offset compensation factor σ, which needs to be calibrated. It is

possible to decrease the SINR level of the best users and increase the SINR of the worst users

to different values of σ. Table 4.5 shows the SINR values of the of 5th and 95th percentiles

to both communications, and the difference between those percentiles to different values

of σ. Remembering if the value of σ is small, the users can achieve the same SINR level.

Finally, σ = 0.8 is set, because it has the smallest difference between percentiles, as shown in

Table 4.5.

After the choice of parameters, the PC schemes are evaluated. Figure 4.8 shows the CDFs

of the SINR values for cellular or D2D links, as it can be seen in Figure 4.8(a), the behavior

of SINR levels of cellular links, when the same PC scheme is applied to both communications.

Comparing OLPC and CLPC it is possible to perceive that CLPC improves the worst users

without compromising the best users.

The SDPC modifies the power of users who show SINR between Γmax and Γmin and keeps

a fixed power value given by Equation (3.11) to user’s SINR values below Γmin. Comparing

EPA with both LTE PC schemes and SDPC, a decrease in the SINR level of the users with

4.1. Power Control 34

Table 4.5: Calibration of σ for CLPC

σ SINR5% SINR95% SINR95% − SINR5%

Cellular Communication

0.2 -26.48 -1.34 25.140.4 -21.96 0.16 22.120.6 -20.45 1.17 21.620.8 -15.92 3.18 19.101.0 -19.94 9.21 29.10

D2D Communication

0.2 -13.41 8.70 22.110.4 -10.39 8.21 18.600.6 -9.38 8.21 17.590.8 -7.87 9.20 17.071.0 -10.00 10.73 20.73

high SINR level can be seen, and this behavior provides a better power efficiency to cellular

communication.

Figure 4.8(b) presents SINR levels of D2D links, when PC scheme is applied to both

communications. It can be noted that the SINR has the worst level when OLPC is applied

in D2D links. Special attention must be given to the OLPC and CLPC, since there is a fall

of SINR level when OLPC is used. This fall is due to the high path gain caused by proximity

between D2D transmitter and receiver; however, CLPC is not affected, because there is a

feedback that adjusts transmit power levels. The SDPC keeps the same SINR to users with

low SINR level and improves the users with high SINR level in relation the CLPC.

−30 −20 −10 0 10 20 30 40 500

20

40

60

80

100

CD

F (

%)

SINR (dB)

EPA Cell

SDPC Cell

OLPC Cell

CLPC Cell

(a) SINR of cellular communications.

−30 −20 −10 0 10 20 30 40 500

20

40

60

80

100

CD

F (

%)

SINR (dB)

EPA D2D

SDPC D2D

OLPC D2D

CLPC D2D

(b) SINR of D2D communications.

Figure 4.8: SINR by applying power control schemes to cellular and D2D links.

To further analyze the performance of PC schemes, Figure 4.9 shows the behavior of PC

schemes in relation to spectral and power efficiency. PC schemes with high spectral efficiency

are situated at the top of figure and high power efficiency are situated in the right of the figure.

From a cellular communications point of view, it is possible to note that EPA has the highest

spectral efficiency and the lowest power efficiency among the studied PC schemes, because

it always uses high transmit power. CLPC and OLPC have about the same power efficiency,

however, CLPC has a feedback, which increases its spectral efficiency. Both LTE PC schemes

have a spectral efficiency higher than SDPC, because SDPC provides a balance between

spectral and power efficiency. So that the SDPC keeps a reasonable spectral efficiency and

provides a gain of 70% in power efficiency compared with LTE PC schemes.

Considering D2D communications, EPA keeps the same behavior of the cellular

4.1. Power Control 35

communications. Both LTE PC schemes have low level of spectral efficiency, however, OLPC

has a little more decreased spectral efficiency, achieving the highest power efficiency. When

SDPC and LTE PC schemes are compared, it is possible to note that SDPC shows better

spectral efficiency. However, when power efficiency is compared, SDPC has a gain of 35% in

relation to CLPC and a loss of 120% in relation to OLPC. Another result that can be noted

is that PC schemes in the middle of the Figure 4.9, it can be combined to provide a tradeoff

between spectral efficiency and power efficiency.

5 10 15 20 25 30

0.3

0.8

1.3

1.8

SDPC

OLPC

CLPC

EPA

SDPC

OLPC

CLPC

EPA

Power efficiency [bps/Hz/cell/W]

Spec

tral

effi

cien

cy[b

ps/

Hz/

cell

]Cellular comm.D2D comm.

Figure 4.9: Performance of PC schemes for cellular and D2D communications.

4.1.2.2 Closed Loop Soft Dropping (CLSD) a hybrid PC scheme

The CLSD is a hybrid PC scheme based on CLPC and SDPC. For performance evaluation,

CLSD parameters are set with the values that have provided a good performance to CLPC

and SDPC in their original form in the Section 4.1.2.1. The parameters are summarized in

Table 4.6.

Table 4.6: CLSD parameters

Parameter Cellular D2D

α 0.3 0.5β 0.8 0.8∆P 10 dB 20 dBΓmax 25 dB 25 dBΓmin 20 dB −5 dB

β and σ have the same function

Figure 4.10 shows the results obtained in terms of spectral efficiency. CLSD provides the

best results of total spectral efficiency, due to knowledge of path gain, current SINR and to be

able of modifing target SINR.

Another way to view results of Figure 4.10 is in terms of relative gains. Table 4.7

summarizes spectral efficiency relative gains when CLSD is compared with other algorithms.

CLSD provides a reasonable performance to cellular communication with the highest and

lowest relative gain are of 91% and 22% compared with SDPC and EPA, respectively. From

a D2D communication point of view, CLSD has a reasonable spectral efficiency gains with

the highest and lowest relative gains are 275% and −7% compared with OLPC and EPA,

respectively.

Figure 4.11 determines the power efficiency of the PC schemes. CLSD achieves

18 bps/Hz/cell/W for cellular communications, which is the best result among all studied PC

schemes, while EPA has the lowest power efficiency achieving 7 bps/Hz/cell/W. In other words,

CLSD manages smartly the power transmit and EPA wastes it, because the transmit power is

high for all users when EPA is used. From D2D point of view, CLSD provides the second best

4.1. Power Control 36

EPA OLPC CLPC SDPC CLSD0

0.5

1

1.5

2

2.5

3

3.5

4

Sys

tem

spe

ctra

l effi

cien

cy [b

ps/H

z/ce

ll]

CellularD2DTotal

Figure 4.10: Spectral efficiency of PC schemes for cellular and D2D communications.

Table 4.7: Spectral efficiency relative gains applying CLSD compared with other PC schemes (%).

EPA OLPC CLPC SDPC

Cellular links 22 76 69 91D2D links -7 275 114 53

Total 8 124 85 73

D2D power efficiency. The reason of this high power efficiency for OLPC is the path gain of

D2D communications described in Section 4.1.2.1.

The power efficiency relative gains of CLSD compared with other PC schemes are described

in Table 4.8. It is important to highlight the highest relative gain to cellular and D2D

communications, which are 157% and 100% when compared with EPA.

It is important to highlight that the CLSD has this good performance due to knowledge of

path gain, current SINR and to be able of modify target SINR. These information are useful

to improve spectral and power efficiency of the system, however, the complexity of CLSD and

the number of subcarriers used to feedback is higher compared with other PC schemes.

EPA OLPC CLPC SDPC CLSD0

5

10

15

20

25

30

Pow

er e

ffici

ency

[bps

/Hz/

cell/

W]

CellularD2D

Figure 4.11: Power efficiency of PC schemes for cellular and D2D communications.

4.1. Power Control 37

Table 4.8: Power efficiency relative gains applying CLSD compared with other PC schemes (%).

EPA OLPC CLPC SDPC

Cellular links 157 80 75 29D2D links 100 -59 41 9

4.1.2.3 Impact of loads in PC schemes

The analysis of PC schemes in a scenario with different loads is important to understand

if they explore well the diversity that each user provides in the system. In the simulations,

both communications use the same PC scheme and its total spectral and power efficiency are

evaluated. It is seen in Figure 4.12 that EPA achieves good spectral efficiency when the offered

load increases, it surpass SDPC, OLPC and CLPC, however, this efficiency range decreases for

high loads, because EPA does not explore well the diversity that each user provides.

When SDPC is used in both communications, it achieves better results than OLPC and

CLPC, because it has feedback and variable target SINR. These two features offer the

opportunity of increasing the SINR of the worst users and keep reasonable SINR levels for

the best users.

Taking a look at LTE PC schemes, it is perceptible that the OLPC and CLPC have a similar

behavior. However, OLPC has a marginal loss due to the lack of feedback, which is present in

CLPC. Finally, the CLSD has achieved the best performance for all considered loads, because

it uses the benefits of both CLPC and SDPC.

4 8 16 321

2

3

4

5

6

7

Load (Number of Users)

Tot

al s

pect

ral e

ff. [b

ps/H

z/ce

ll]

EPAOLPCCLPCSDPCCLSD

Figure 4.12: Total spectral efficiency comparison for different loads.

From Figure 4.13(a), one sees that EPA shows the worst result in terms of power efficiency

to cellular communications. This is an indication that EPA fails in explore the diversity of

users, given that it always uses the maximum transmit power regardless of user SINR.

The OLPC and CLPC provide a similar power efficiency for four users in each cell. However,

CLPC for high loads attains a significant power efficiency gain compared with OLPC. This

results show that only knowledge of path gain is not enough to provide a good power efficiency

to cellular communications, because it shows a low information about user in the network

to PC scheme. In order to offer better power efficiency to cellular communications SDPC and

CLSD are the best choices, which achieve good performance in a scenario with high loads due

to explore well the diversity.

It may be seen in Figure 4.13(b) that incorporating EPA, D2D communications do not show

good performance in terms of power efficiency. It is possible to note that the power efficiency

decreases after 8 users in a cell, because the interference level is so high that harms the

4.1. Power Control 38

spectral efficiency of D2D users. The OLPC keeps a good performance for all offered loads.

This behavior can be explained by the high value of the path gain due to the proximity of

communications occurring inside the hotspot.

It is interesting to see that CLPC shows a low power efficiency compared with OLPC,

because it tries to keep a good spectral efficiency, therefore, it does not decrease the transmit

power as much as OLPC. The SDPC and CLSD for low loads have a similar power efficiency

for D2D communications, however, CLSD is more efficiency for high loads.

The reason for the power efficiency gain of CLSD is that it has not only information about

the path gain, which decreases the power transmit like OLPC, but also it has variable target

SINR that provides a good total spectral efficiency.

4 8 16 320

10

20

30

40

50

60

70

80

Load (Number of Users)

Pow

er e

ffici

ency

[bps

/Hz/

cell/

W]

EPAOLPCCLPCSDPCCLSD

(a) Power efficiency for cellular communications.

4 8 16 320

50

100

150

Load (Number of Users)

Pow

er e

ffici

ency

[bps

/Hz/

cell/

W]

EPAOLPCCLPCSDPCCLSD

(b) Power efficiency for D2D communications.

Figure 4.13: Power efficiency comparison for different loads in cellular and D2D communications.

4.1.2.4 Imperfect CSI

Features such as multi-user scheduling operating in fading channels can be used to

explore diversity gains and improve the quality of communications in cellular networks. For

this purpose, precise CSI (i.e. perfect CSI) needs to be available at the eNB to perform rate

adaptation and scheduling. However, in real cellular networks, CSI is impaired by channel

estimation errors and feedback delays. This impact is high in networks where the data is sent

to a central point, because high backhaul latency can cause CSI imperfections, resulting in

performance degradations [75,76].

In order to understand the effects of imperfect CSI in D2D communications underlaying

cellular networks, a scenario where both communications use the same PC scheme with the

parameters described in Section 4.1.2.1 and Section 4.1.2.2 is used. Therein, each site has

16 UEs operating in UL. The reason for choosing this scenario is due to the interference level

and effects of delay feedback being significant.

Figure 4.14 shows the total spectral efficiency to different delays ranging from 0 TTI (no

delay) to 5 TTIs. It is noticeable that without feedback delays the CLSD has the best spectral

efficiency, followed by EPA, SDPC, CLPC and OLPC. All PC schemes decrease its spectral

efficiency when delay increases, however, each PC scheme has a different drop rate. The EPA

and OLPC have slight loss of spectral efficiency compared with other PC schemes, because

EPA and OLPC are not influenced significantly by feedback. In other words, EPA does not

need feedback, because it always uses the same transmit power and OLPC requires only G

(path gain), which does not suffer a significant modification from one Transmission Time

4.1. Power Control 39

Interval (TTI) to another. The main factor that harms EPA and OLPC is scheduling, because

eNB allocates PRB to users, which decrease their quality of channel from one TTI to another.

The CLPC has an accentuated loss of spectral efficiency compared with EPA and OLPC,

because CLPC computes transmit power based on G (path gain) and current SINR, so CLPC

computes transmit power using two out-of-date measures.

The SDPC and CLSD have the worst spectral efficiency for high delays, because SDPC is

dependent on the current SINR, target SINR and previous transmit power. Moreover, CLSD is

not only dependent on same parameters as SDPC measures, but also on CLPC measures.

The PC schemes have a similar spectral efficiency when the delay increases up to one

TTI, the spectral efficiency keeps between 4.8 bps/Hz/cell and 5.4 bps/Hz/cell. This difference

becomes expressive when delay is higher than one TTI, in this case the difference between

values achieves 3.4 bps/Hz/cell when delay is 5 TTIs.

0 1 2 3 4 50

1

2

3

4

5

6

Delay (TTI)

Tot

al s

pect

ral e

ff. [b

ps/H

z/ce

ll]

EPAOLPCCLPCSDPCCLSD

Figure 4.14: Total spectral efficiency for different delays.

As it is shown in Figure 4.15(a), cellular power efficiency has the same behavior of spectral

efficiency, that is to say, power efficiency decreases for high delay. SDPC and CLSD have the

best results in terms of power efficiency when CSI is perfect, however, these PC schemes are

affected negatively after a delay of 2 TTIs. CLPC has an acceptable power efficiency for low

delay values, however, it is outweighed by OLPC for high delay. EPA has the worst results in

terms of power efficiency up to a delay of 4 TTI and after this delay value, CLSD has the worst

performance due to the number of out-of-date measurements.

Figure 4.15(b) presents D2D power efficiency, it is noticeable that OLPC has the lowest loss

of power efficiency compared with other PC schemes. This behavior is due to G (path gain) of

D2D communications to be similar from one TTI to another, due to the proximity among UEs

communicating in D2D mode. Among the PC schemes, SDPC and CLSD have a high power

efficiency loss, while CLPC keeps a reasonable performance.

The PC schemes based on many measures suffer a significant loss of spectral and power

efficiency, when feedback delay occurs. In terms of total spectral efficiency and cellular power

efficiency, feedback delay becomes significant when the system has a delay higher than 2

TTIs, thus it is better to use PC schemes simpler to provide the best efficiency to the system.

It is interesting to note that OLPC keeps a good power efficiency to D2D communications

independent of the delay, because OLPC provides transmit power based on the metric G (path

gain), which does not vary significantly among TTIs.

4.1. Power Control 40

0 1 2 3 4 50

10

20

30

40

50

Delay (TTI)

Pow

er e

ffici

ency

[bps

/Hz/

cell/

W]

EPAOLPCCLPCSDPCCLSD

(a) Power efficiency for cellular communications.

0 1 2 3 4 50

20

40

60

80

100

Delay (TTI)

Pow

er e

ffici

ency

[bps

/Hz/

cell/

W]

EPAOLPCCLPCSDPCCLSD

(b) Power efficiency for D2D communications.

Figure 4.15: Power efficiency for different delays in cellular and D2D communications.

4.1.2.5 Convergence of SD

It is important to verify the convergence of SD, an analytical analysis of convergence

is demonstrated in Appendix A. In this master’s thesis, a complementary analysis using

computational simulations is also realized. It is important to make clear that the convergence

is influenced by parameters values, loads, cell size and CSI. So, a scenario with 8 users,

using SD parameters ∆P = 30 dB, Γmax = 25 dB and Γmin = −5 dB was selected for analysis

considering both cellular and D2D communications.

In order to show the convergence of SD, it was used Mean Squared Error (MSE), which is

based on values of target SINR Γk,n and current SINR γk,n of user k scheduled to PRB n in each

cell site. In Figure 4.16, I illustrate the step-by-step on how the convergence is calculated.

First step is to create a matrix that contains the difference between target SINR Γk,n and user

SINR γk,n squared. Second step is to create another matrix, where each element corresponds

the mean of difference between target SINR Γk,n and user SINR γk,n squared to each TTI. Third

step represents the mean to each sample. Finally, the last step aims at the normalization of

the elements.

. .

Sites

PRBs TTIs

Sam

ples

. . .

1 . . .

( )γΓ nk,nk,2 −

Mean

Normalised

Mean

∑∑ −= =⋅

N Nk PRB

k nSitePRB NN 1 1

2

)γΓ( nk,nk,1

(1) (2)

(3)

(4)

.

Figure 4.16: Detailed description of calculation of convergence.

Figure 4.17(a) shows the MSE for the SINR of cellular and D2D communications, the mean

of difference between target SINR Γk,n and user SINR γk,n squared. It shows that SD keep a

decrease from 153 to 5 , this means that the variation of the difference was about 10 dB, in

other words, the SDPC improved the accuracy of target SINR in 10 dB.

Figure 4.17(b) shows the convergence using normalization of cellular communications in

DL/UL and D2D communications, respectively. In both communications, the Normalized

4.2. Antenna Downtilt 41

Mean Square Error (NMSE) achieves values lower than 0, 1 in 3TTIs (3ms).

Figure 4.17(c) shows the behavior of convergence to 200TTIs (200ms). It is interesting to

note that some peaks appear in the figure. These peaks are due to MaxGain scheduling, which

modifies the PRB allocation during the simulations, thus changing channel parameters. Even

though, SD is able to return the normal operation after 3 or 4TTIs (3ms or 4ms).

0 50 100 150 2000

20

40

60

80

100

120

140

153

TTI

Mea

n sq

uare

err

or

Cellular DLCellular ULD2D

(a) Convergence during 200TTIs using MSE.

2.9 2.95 3 3.05 3.10

0.02

0.04

0.06

0.08

0.1

TTI

Nor

mal

ized

mea

n sq

uare

err

or

Cellular DLCellular ULD2D

(b) Convergence to first values of TTIs using NMSE.

50 100 150 2000

0.2

0.4

0.6

0.8

1

TTI

Nor

mal

ized

mea

n sq

uare

err

or

Cellular DLCellular ULD2D

(c) Convergence during 200TTIs using NMSE.

Figure 4.17: Convergence of SD.

4.2 Antenna Downtilt

This section investigates the impact of electrical downtilt in an urban-macrocell scenario

where D2D communications underlay a cellular network. The downtilt evaluation is realized

in the DL through system-level simulations, which are aligned with the LTE architecture [60,

63,64,66].

4.2.1 Impact of Downtilt in a cellular network with D2D

For the performance evaluation, the scenario detailed in Table 2.3 is used. Initially, only

EPA is utilized to determine the power of the UEs. This strategy is useful to better understand

the behavior of downtilt in a scenario with D2D communications, because in this scenario the

gains are not influenced by PC scheme.

In Figure 4.18 shows the power and spectral efficiency values as functions of the downtilt

angle. It is possible to note that cellular spectral efficiency has good values when the downtilt

angle increases up to 12° (arrow number 1 ) since cell isolation is improved and interference

is reduced when the downtilt angle increases (Figure 4.18(a)). However, after this angle the

cellular users performance begins to decay, since too large downtilt angles reduced coverage

(arrow number 2 ).

4.2. Antenna Downtilt 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 170.5

1

1.5

2

2.5

3

3.5

4

Angle (°)

Cel

lula

r sp

ectr

al e

ff. [b

ps/

Hz/

cell

]

2

1

(a) Cellular spectral efficiency.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 170.2

0.4

0.6

0.8

1

1.2

1.4

Angle (°)

D2D

spec

tral

eff

. [b

ps/

Hz/

cell

]

3

4

(b) D2D spectral efficiency.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 171

1.5

2

2.5

3

3.5

4

4.5

Angle (°)

Tota

l sp

ectr

al e

ff. [b

ps/

Hz/

cell

]

(c) Total spectral efficiency.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 170

1

2

3

4

5

6

Angle (°)

Pow

er e

ffic

iency

[bps/

Hz/

cell

/W]

Cellular Communications

D2D Communications

(d) Power efficiency.

Figure 4.18: Behavior of spectral and power efficiency for different levels of tilt.

From a D2D point of view (Figure 4.18(b)), spectral efficiency has the opposite behavior

compared with cellular UEs. At first, a spectral efficiency decline occurs, which reaches the

worst result at 7° (arrow number 3 ), since the Evolved Node B (eNB) interference is focused in

the hotspot area near the cell edge. After 7°, it is possible to improve D2D spectral efficiency,

because the interference inside hotspot zone decreases (arrow number 4 ).

Figure 4.18(c) shows the total spectral efficiency to downtilt angles between 1° and 17°.

The rectangle determines the range of downtilt angles where the total spectral efficiency is

higher than without downtilt (0°) and does not harm the coverage area of cellular users. In

other words, angles between 8° and 12° improve spectral efficiency of system, while preserving

cellular communications.

In the following, it is adopted an angle of 12° for downtilt in the simulations, since it

corresponds to the angle that provided the best total spectral efficiency in the previous

evaluation. Then, the influence of the downtilt angle in the SINR and interference curves

is analyzed, as illustrated in Figure 4.19. It is shown in Figures 4.19(a) that using downtilt

it is possible to improve the SINR curves for cellular and D2D communications. The analysis

at 50% in Figure 4.19(b) confirms the results of the SINR curves, because the interference of

cellular and D2D communications decreases 10 dB and 4 dB, respectively.

Figure 4.20 compares the system spectral efficiency of cellular and D2D communications

applying 12° as downtilt angle at eNB.

D2D communications do not get a significant gain in spectral efficiency, while cellular

communications get a performance which is better than in the conventional scenario, where

0° of downtilt is used. The total spectral efficiency achieves 58% of gain, so that it can be

concluded that angles between 8° and 12° offer the possibility to keep D2D communications’

4.2. Antenna Downtilt 43

−30 −25 −20 −15 −10 −5 0 5 10 15 20 25 30 35 400

10

20

30

40

50

60

70

80

90

100

CD

F (

%)

SINR (dB)

0 ° − Cell users

0 ° − D2D users

12 ° − Cell users

12 ° − D2D users

(a) SINR Curves.

−95 −90 −85 −80 −75 −70 −65 −60 −55 −500

10

20

30

40

50

60

70

80

90

100

CD

F (

%)

Interference power (dBm)

0 ° − Cell users

0 ° − D2D users

12 ° − Cell users

12 ° − D2D users

(b) Interference Curves.

Figure 4.19: SINR and interference levels by applying downtilt.

quality while improving considerably the performance of cellular communications. In terms

of power efficiency, the behavior is shown in Figure 4.18(d) and the values are summarized in

Table 4.9. It is possible to note that angles between 11° and 12° have better power efficiency to

both communications, while angles lower than 11° do not have good power efficiency for D2D

communications.

Table 4.9: Power efficiency relative gains for different downtilt angles compared without downtilt (%).

8◦ 9◦ 10◦ 11◦ 12◦

Cellular communications 4% 22% 40% 53% 65%D2D communications -29% -23% -0.1% 1.7% 22%

Cellular D2D Total0

1

2

3

4

5

Syst

em s

pec

tral

eff

icie

ncy

[bps/

Hz/

cell

]

12°

Figure 4.20: System spectral efficiency of cellular and D2D communications in scenario with andwithout downtilt.

In Figure 4.21, the impact of outage to both communications and the outage reduction

compared without downtilt can be seen. This outage reduction represents the number of

users who previously were in outage and currently are not in outage. For example, 10 users

are in outage without downtilt and an outage reduction of 50% with downtilt means that 5 of

the 10 users are not in outage currently.

The positive impact of downtilt occurs in a range of angles between 7° and 15° which are

4.2. Antenna Downtilt 44

named critical angles because angles lower than 7° and higher than 15° do not reduce the

outage to cellular communications. Angles out of this range must be avoided.

Analyzing the cellular communications, it is possible to note that for angles from 7° to 12°

a reduction of the outage level occurs due to decrease of the intercellular interference level.

After 12°, the coverage radius of the eNB reduces to each angle, leaving cellular users without

communication.

From the D2D communication point of view, the outage reduction decreases for low

downtilt angles. However, it keeps a gain in relation to case without downtilt. After 7° a

decrease of the outage level occurs because the interference level inside the hotspot decreases.

It is important to clarify that the focus is to provide the best outage level to cellular users.

When 12° of downtilt is used, it is possible to achieve an outage reduction of 75% for cellular

communications, while D2D communications achieve a gain of 28%.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 170

10

20

30

40

50

60

70

80

Angle (°)

Out

age

redu

ctio

n (%

)

Cellular communicationsD2D communications

Critical angle

Figure 4.21: Outage reduction for different levels of tilt.

4.2.2 SDPC in a Downtilt scenario

Section 4.2.1 shows the impact of downtilt in a scenario only with EPA. In this section,

the performance of SDPC is studied jointly with downtilt. The SDPC is evaluated using 12°of

downtilt, because this angle achieved good results in terms of spectral and power efficiency

with EPA.

Figure 4.22 shows the total spectral efficiency and power efficiency, when cellular and

D2D communications use SDPC and EPA, respectively. Figure 4.22(a) shows the case where

SDPC does not use downtilt. When a downtilt of 12° is used (see Figure 4.22(b)) the spectral

efficiency increases for all sets of parameters analyzed. The highest possible spectral efficiency

(∆P = 40 dB and Γmin = 20 dB) using SDPC without downtilt is 3 bps/Hz/cell, however, it is

possible to achieve 4.5 bps/Hz/cell using 12° of tilt, in other words, SDPC working together with

downtilt provide a gain of 50%.

Power efficiency has a similar behavior of spectral efficiency, because it increases for all

set of parameters due to cell isolation. Taking a look in Figure 4.22(c), it is evident that

the parameters ∆P = 20 dB and Γmin = −5 dB provide the highest power efficiency, the gain

achieved in this point using downtilt (see Figure 4.22(d)) is 63% compared without downtilt.

Figure 4.23 shows the total spectral efficiency and power efficiency, when cellular and D2D

communications use EPA and SDPC, respectively.

Taking a look in Figure 4.23(a), it is possible to note that spectral efficiency has small

variation when the parameters are modified, however, after the eNB changes downtilt to 12° as

4.2. Antenna Downtilt 45

010

2030

40

−50

510

1520

2

2.5

3

3.5

4

4.5

PC range [dB]Min. target SINR [dB]

Tot

al s

pect

ral e

ff. [b

ps/H

z/ce

ll]

(a) Behavior of spectral efficiency applying 0°of tilt.

010

2030

40

−50

510

1520

3

3.5

4

4.5

PC range [dB]Min. target SINR [dB]

Tot

al s

pect

ral e

ff. [b

ps/H

z/ce

ll]

(b) Behavior of spectral efficiency applying 12°of tilt.

010

2030

40

−50

510

1520

0.5

1

1.5

PC range [dB]Min. target SINR [dB]

Pow

er e

ffici

ency

[bps

/Hz/

cell/

W]

(c) Behavior of power efficiency applying 0°of tilt.

010

2030

40

−50

510

1520

0.8

1

1.2

1.4

PC range [dB]Min. target SINR [dB]

Pow

er e

ffici

ency

[bps

/Hz/

cell/

W]

(d) Behavior of power efficiency applying 12°of tilt.

Figure 4.22: Total spectral efficiency and power efficiency (SDPC in cellular and No-PC in D2D links).

shown in Figure 4.23(b), it is possible to check a higher diversity than in the previous results.

The highest value of spectral efficiency without downtilt is 2.8 bps/Hz/cell, when ∆P = 40 dB

and Γmin = 20 dB, while using a downtilt angle of 12° it is possible to provide 4.4 bps/Hz/cell/W,

which represents a gain of 57%.

Finally, Figures 4.23(c) and 4.23(d) show how much beneficial is downtilt and SDPC

working together. The power efficiency gain reaches 80% compared with the case without

downtilt, using ∆P = 20 dB and Γmin = −5 dB. This is a clear evidence that the downtilt

decreases inter-cell interference for D2D links. When interference level becomes low due to

downtilt, the SDPC decreases the transmit powers, thus increasing energy efficiency.

The Figures 4.22 and 4.23 show that downtilt provides the opportunity of the SDPC to

further improve the performance of both communications, since the interference level is

reduced due to downtilt, the SDPC can provide high target SINR, while it saves transmit

power of eNB and UE achieving a better power efficiency.

This is indicative that the downtilt working together with the PC schemes can provide high

gains of power and spectral efficiency not only in conventional network, but also when D2D

communications underlaying cellular networks.

4.2. Antenna Downtilt 46

010

2030

40

−50

510

1520

3

3.5

4

4.5

PC range [dB]Min. target SINR [dB]

Tot

al s

pect

ral e

ff. [b

ps/H

z/ce

ll]

(a) Behavior of spectral efficiency applying 0° of tilt.

010

2030

40

−50

510

1520

4.2

4.3

4.4

4.5

PC range [dB]Min. target SINR [dB]

Tot

al s

pect

ral e

ff. [b

ps/H

z/ce

ll]

(b) Behavior of spectral efficiency applying 12° of tilt.

010

2030

40

−50

510

1520

2.5

3

3.5

4

4.5

PC range [dB]Min. target SINR [dB]

Pow

er e

ffici

ency

[bps

/Hz/

cell/

W]

(c) Behavior of power efficiency applying 0°of tilt.

010

2030

40

−50

510

15202.5

3

3.5

4

4.5

PC range [dB]Min. target SINR [dB]

Pow

er e

ffici

ency

[bps

/Hz/

cell/

W]

(d) Behavior of power efficiency applying 12° of tilt.

Figure 4.23: Total spectral efficiency (No-PC in cellular and SDPC in D2D links).

Chapter 5Conclusions

This master’s thesis has dealt with Radio Resource Management (RRM) for cellular and

network-assisted Device-to-Device (D2D) communications, as well as strategies used to

improve energy efficiency in a scenario where D2D communications underlays the cellular

network. Strategies for interference management, such as Power Control (PC) and downtilt

have been analyzed and calibrated seeking for the minimum waste of energy in the cellular

network without harming its capacity.

This master’s thesis addressed the benefits of the Soft Dropping (SD) algorithm for cellular

and D2D communications in the Downlink (DL) of a multi-cell scenario through system-level

simulations. Results indicate that the SD algorithm in a Micro-cell scenario is effective in

controlling the trade-off between system spectral efficiency of cellular communications and

power saving of D2D transmitters. The application of SD in D2D communications always

provides better performance in terms of total system spectral efficiency and power economy

for any operation point than the application of SD to cellular communications. The main

reason for that is related with the reduction of the high interference power originated from

D2D communications. In DL, D2D transmitters act as interfering sources close to cellular

receivers while D2D receivers are far way from both cellular and D2D transmitters. Thus, the

SD algorithm appears as a promising solution to protect cellular communications from the

interference caused by D2D communications.

PC schemes to protect cellular communications and achieve higher energy efficiency gains

for D2D communications on the Uplink (UL) in a Micro-cell scenario were also investigated.

Results indicate that in terms of energy efficiency the Soft Dropping Power Control (SDPC)

performs better for cellular links while the Open Loop Power Control (OLPC) provides high

gains for D2D links. While high values of α have been widely used in OLPC studies,

α ∈ {0.4, 0.5} has provided high energy efficiency gains for D2D links. Considering the most

favorable scenario for sharing resources in all cells, it was also seen that the minimum cost

for enabling system spectral efficiency gains for D2D communications represents a minimal

impact of 11% on the system spectral efficiency of cellular communications.

Indeed, different PC schemes vary greatly in complexity, numbers of parameters, and

have different performance levels. It has been noted that the Equal Power Allocation (EPA)

scheme always has the highest spectral efficiency and the lowest power efficiency in both

communications. SDPC keeps a reasonable spectral efficiency and provides a gain of 70%

in power efficiency compared to the Long Term Evolution (LTE) PC schemes for cellular

communications. If the purpose of PC is to be power efficient, it would be interesting to

48

use SDPC in cellular communications and OLPC in D2D communications.

We also conclude that for OLPC and Closed Loop Power Control (CLPC), path gain is an

important factor affecting performance of the both communication modes and the factor σ =

0.8 can modify the behavior of CLPC, because it increases the Signal to Interference-plus-Noise

Ratio (SINR) of the worst users.

Another conclusion is that Closed Loop Soft Dropping (CLSD), which is based on SDPC

and CLPC provides a tradeoff between spectral efficiency and power efficiency, because CLSD

has good performance due to knowledge of path gain, current SINR and because it is able to

modify the target SINR values. These information are useful to improve spectral and power

efficiency of the system, however, the complexity of CLSD and the number of subcarriers used

to feedback is higher compared with other PC schemes.

It was shown that algorithms PC schemes are influenced by the load of system. PC

schemes such as CLSD provide the best results in terms of total spectral efficiency when

the load increases, because it uses the benefits of both CLPC and SDPC, such as feedback

and variable target SINR. In terms of power efficiency, EPA shows the worst result to both

communications, while SDPC and CLSD provide good results to cellular communications

due to explore the diversity. From D2D point of view, OLPC keeps a good performance for

all offered loads. This behavior can be explained by the high path gain values due to the

proximity of communications inside the hotspot.

In terms of Channel State Information (CSI), we concluded that PC schemes based on many

measures suffer a significant loss of spectral and power efficiency when subject to feedback

delay. High delays harm the total spectral efficiency and cellular power efficiency, so in this

case it should be adopted simple PC schemes to provide the best efficiency to the system.

From the D2D point of view, OLPC keeps a good power efficiency independent of the delay,

because OLPC provides transmit power based on the metric G (path gain), which does not

vary significantly among Transmission Time Intervals (TTIs).

Based on the results of this master’s thesis, we concluded that antenna downtilt can be

used as a simple and efficient technique not only in conventional cellular networks, but

also for D2D communications underlying cellular networks. Regions that offer improved

spectral and power efficiency to both types of communication were determined and the range

between 8° and 12° for downtilt angle provided good spectral efficiency to cellular and D2D

communications. However, angles between 8° and 10° are not good parameter values for D2D

communications, since they do not lead to power efficiency gains. In this way, angles between

11◦ and 12◦ should be chosen, which provide gains in terms of power efficiency to both cellular

and D2D communications that reach 65% and 22%, respectively, while they improved the total

spectral efficiency in 58%.

Downtilt working together with SDPC schemes brings opportunity to reduce inter-cell

interference in D2D links due to downtilt and to save transmit power due to SDPC. In other

words, downtilt intensifies the gain of SDPC.

This master’s thesis intends to contribute to a better understanding of the role and

behavior of PC schemes when D2D communications underlay a cellular network.

Appendix AProof of convergence SDPC

The target SINR Γk,c,n(p(t)k,c,n) of User Equipment (UE) k in the cell c and Physical Resource

Block (PRB) n at TTI t is given according to

Γk,c,n(p(t)k,c,n) =

Γmax, p(t)k,c,n ≤ Pmin,

Γmax

(

p(t)k,c,n

Pmin

, Pmin < p(t)k,c,n < Pmax,

Γmin, p(t)k,c,n ≥ Pmax,

(A.1)

where

ρ =log10(Γmin/Γmax)

log10(Pmax/Pmin). (A.2)

Then, the power per PRB of each UE is updated every transmission as follows

p(t+1)k,c,n = p

(t)k,c,n

(

Γk,c,n(p(t)k,c,n)

γk,c,n(p(t))

, (A.3)

By assuming η as the thermal noise at the receiver and tx(m) the transmitter D2D pair

m ∈ {0, 1, . . . , R}, the SINR (γ(t)k,c,n) perceived by of cellular user k in the cell c and PRB n at TTI

t can be written as show in Equation (A.4):

γ(t)k,c,n =

∣∣∣h

(t)k,c,n

∣∣∣

2

p(t)k,c,n

C∑

c′ 6=c

k∑

k′

∣∣∣h

(t)k,c′,n

∣∣∣

2

p(t)k′,c′,n

︸ ︷︷ ︸

Interference from cellular links

+C∑

c′

M∑

m′

∣∣∣h

(t)k,tx(m′),c′,n

∣∣∣

2

p(t)tx(m′),c′,n

︸ ︷︷ ︸

Interference from D2D links

+η2

(A.4)

50

For power value in Pmin < p(t)k,c,n < Pmax

I(p(t)k,n

) = p

(t+1)k,c,n

= p

(t)k,c,n

Γmax

p

(t)k,c,n

Pmin

ρ

∣h(t)k,c,n

2p(t)k,c,n

C∑

c′ 6=c

k∑

k′

∣h(t)

k,c′,n

2p(t)

k′,c′,n+

C∑

c′

M∑

m′

h(t)

k,tx(m′),c′,n

2

p(t)

tx(m′),c′,n+η2

β

,

= p

(t)k,c,n

Γmax

p

(t)k,c,n

Pmin

ρ(C∑

c′ 6=c

k∑

k′

∣h

(t)k,c′,n

2p

(t)k′,c′,n

+

C∑

c′

M∑

m′

∣h

(t)k,tx(m′),c′,n

2p

(t)tx(m′),c′,n

+ η

2

)

∣h

(t)k,c,n

2p

(t)k,c,n

β

,

= p

(t)k,c,n

(

p

(t)k,c,n

)ρβ

(

p

(t)k,c,n

Γmax

(

C∑

c′ 6=c

k∑

k′

∣h

(t)k,c′,n

2p

(t)k′,c′,n

+

C∑

c′

M∑

m′

∣h

(t)k,tx(m′),c′,n

2p

(t)tx(m′),c′,n

+ η

2

)

∣h

(t)k,c,n

2P

ρmin

β

,

=

(

p

(t)k,c,n

)1+ρβ−β

Γmax

(

C∑

c′ 6=c

k∑

k′

∣h

(t)k,c′,n

2p

(t)k′,c′,n

+

C∑

c′

M∑

m′

∣h

(t)k,tx(m′),c′,n

2p

(t)tx(m′),c′,n

+ η

2

)

∣h

(t)k,c,n

2P

ρmin

β

,

(A.5)

Once the all terms in Equation (A.5) are positive, I(p(t)k,n) satisfies positivity. To verify

monotonicity, it is necessary to ensure that I(p(t)k,c,n) ≥ I(p′

(t)k,c,n), for all p

(t)k,c,n ≥ p′

(t)k,c,n, then

the value of exponent must be positive.

(

p(t)k,c,n

)1+ρβ−β

≥(

p′(t)k,c,n

)1+ρβ−β

,

1 + ρβ − β ≥ 0,

1 + β(ρ− 1) ≥ 0,

β(ρ− 1) ≥ −1,

β(1− ρ) ≤ 1,

β ≤1

(1 − ρ),

(A.6)

To ensure scalability, aI(p(t)k,c,n) ≥ I(ap(t)k,c,n), for a ≥ 1. This way

a1 ≥ a1+ρβ−β,

1 ≥ 1 + ρβ − β,

0 ≥ ρβ − β,

ρβ − β ≤ 0,

ρβ ≤ β,

ρ ≤ 1,

(A.7)

51

For power value in p(t)k,c,n ≤ Pmin

I(p(t)k,n

) = p

(t+1)k,c,n

= p

(t)k,c,n

Γmax∣

∣h(t)k,c,n

2p(t)k,c,n

C∑

c′ 6=c

k∑

k′

∣h(t)

k,c′,n

2p(t)

k′,c′,n+

C∑

c′

M∑

m′

h(t)

k,tx(m′),c′,n

2

p(t)

tx(m′),c′,n+η2

β

,

= p

(t)k,c,n

Γmax

(

C∑

c′ 6=c

k∑

k′

∣h

(t)k,c′,n

2p

(t)k′,c′,n

+

C∑

c′

M∑

m′

∣h

(t)k,tx(m′),c′,n

2p

(t)tx(m′),c′,n

+ η

2

)

∣h

(t)k,c,n

2p

(t)k,c,n

β

,

=

(

p

(t)k,c,n

)1−β

Γmax

(

C∑

c′ 6=c

k∑

k′

∣h

(t)k,c′,n

2p

(t)k′,c′,n

+

C∑

c′

M∑

m′

∣h

(t)k,tx(m′),c′,n

2p

(t)tx(m′),c′,n

+ η

2

)

∣h

(t)k,c,n

2

β

,

=

(

p

(t)k,c,n

)1−β

Γmax

(

C∑

c′ 6=c

k∑

k′

∣h

(t)k,c′,n

2p

(t)k′,c′,n

+

C∑

c′

M∑

m′

∣h

(t)k,tx(m′),c′,n

2p

(t)tx(m′),c′,n

+ η

2

)

∣h

(t)k,c,n

2

β

,

(A.8)

Once the all terms in Equation (A.8) are positive, I(p(t)k,n) satisfies positivity. To verify

monotonicity, it is necessary to ensure that I(p(t)k,c,n) ≥ I(p′

(t)k,c,n), for all p

(t)k,c,n ≥ p′

(t)k,c,n, then

the value of exponent must be positive.

(

p(t)k,c,n

)1−β

≥(

p(t)k,c,n

)1−β

,

1− β ≥ 0,

1 ≥ β,

β ≤ 1,

(A.9)

To ensure scalability, aI(p(t)k,c,n) ≥ I(ap(t)k,c,n), for a ≥ 1. This way

a1 ≥ a1−β,

1 ≥ 1− β,

β ≥ 0,

(A.10)

The main relations are defined below:

ρ ≤ 1, (A.11)

β ≤1

(1− ρ), (A.12)

β ≤ 1, (A.13)

β ≥ 0, (A.14)

52

Finally, we can combining them

−∞ ≤ ρ ≤ 0, (A.15)

0 ≤ β ≤1

(1− ρ), (A.16)

Bibliography

[1] ITU, “Background on IMT-Advanced , recommendation ITU, tech. rep. doc. IMT-ADV/1„”

Tech. Rep., 2008.

[2] L. Suarez, L. Nuaymi, and J.-M. Bonnin, “An overview and classification of research

approaches in green wireless networks,” EURASIP Journal on Wireless Communications

and Networking, vol. 2012, no. 1, pp. 1–18, 2012.

[3] R. L. Batista, C. F. M. e Silva, J. M. B. da Silva Jr., T. F. Maciel, and F. R. P. Cavalcanti,

“What happens with a proportional fair cellular scheduling when D2D communications

underlay a cellular network?” in IEEE WCNC 2014 - Workshop on Device-to-Device and

Public Safety Communications (WCNC’14 - WDPC Workshop), Istambul, Turkey, Apr. 2014,

pp. 260–265.

[4] G. Fodor and N. Reider, “A distributed power control scheme for cellular network assisted

D2D communications,” pp. 1–6, Dec 2011.

[5] J. Niemelä, T. Isotalo, and J. Lempiäinen, Optimum Antenna Downtilt Angles for

Macrocellular WCDMA Network, vol. 5, Issue 5, pp. 816–827, October 2005.

[6] Y.-D. Lin and Y.-C. Hsu, “Multihop cellular: a new architecture for wireless

communications,” in INFOCOM 2000. Nineteenth Annual Joint Conference of the IEEE

Computer and Communications Societies. Proceedings. IEEE, vol. 3, Mar 2000, pp.

1273–1282 vol.3.

[7] L. Lei, Z. Zhong, C. Lin, and X. Shen, “Operator controlled device-to-device

communications in lte-advanced networks,” Wireless Communications, IEEE, vol. 19,

no. 3, pp. 96–104, June 2012.

[8] M. S. Corson, R. Laroia, J. Li, V. Park, T. Richardson, and G. Tsirtsis, “Toward

proximity-aware internetworking,” IEEE Wireless Communications, vol. 17, no. 6, pp.

26–33, Dec. 2010.

[9] Z.-J. Yang, J.-C. Huang, C.-T. Chou, H.-Y. Hsieh, C.-W. Hsu, P.-C. Yeh, and C.-C. Hsu,

“Peer discovery for device-to-device (d2d) communication in lte-a networks,” in Globecom

Workshops (GC Wkshps), 2013 IEEE, Dec 2013, pp. 665–670.

[10] Q. Du, P. Ren, H. Song, Y. Wang, and L. Sun, “On p2p-share oriented routing over

interference-constrained d2d networks,” in Mobile Ad-hoc and Sensor Networks (MSN),

2014 10th International Conference on, Dec 2014, pp. 138–143.

53

Bibliography 54

[11] A. Le, L. Keller, H. Seferoglu, B. Cici, C. Fragouli, and A. Markopoulou, “Microcast:

Cooperative video streaming using cellular and D2D connections,” CoRR, vol.

abs/1405.3622, 2014. [Online]. Available: http://arxiv.org/abs/1405.3622

[12] R.-S. Cheng, C.-M. Huang, and G.-S. Cheng, “A d2d cooperative relay scheme for

machine-to-machine communication in the lte-a cellular network,” in Information

Networking (ICOIN), 2015 International Conference on, Jan 2015, pp. 153–158.

[13] G. Rigazzi, N. K. Pratas, P. Popovski, and R. Fantacci, “Aggregation and trunking of M2M

traffic via D2D connections,” CoRR, vol. abs/1502.01708, 2015. [Online]. Available:

http://arxiv.org/abs/1502.01708

[14] S. Andreev, O. Galinina, A. Pyattaev, K. Johnsson, and Y. Koucheryavy, “Analyzing

assisted offloading of cellular user sessions onto d2d links in unlicensed bands,” Selected

Areas in Communications, IEEE Journal on, vol. 33, no. 1, pp. 67–80, Jan 2015.

[15] T. Han and N. Ansari, “Offloading mobile traffic via green content broker,” Internet of

Things Journal, IEEE, vol. 1, no. 2, pp. 161–170, April 2014.

[16] A. Asadi, Q. Wang, and V. Mancuso, “A survey on device-to-device communication in

cellular networks,” http://arxiv.org/abs/1310.0720v1, oct 2013.

[17] K. Doppler, M. Rinne, C. Wijting, C. Ribeiro, and K. Hugl, “Device-to-device

communication as an underlay to lte-advanced networks,” Communications Magazine,

IEEE, vol. 47, no. 12, pp. 42–49, Dec 2009.

[18] R. Yin, G. Yu, H. Zhang, Z. Zhang, and G. Li, “Pricing-based interference coordination for

d2d communications in cellular networks,” Wireless Communications, IEEE Transactions

on, vol. 14, no. 3, pp. 1519–1532, March 2015.

[19] S. Xu, H. Wang, T. Chen, Q. Huang, and T. Peng, “Effective interference cancellation

scheme for device-to-device communication underlaying cellular networks,” in Vehicular

Technology Conference Fall (VTC 2010-Fall), 2010 IEEE 72nd, Sept 2010, pp. 1–5.

[20] T. Peng, Q. Lu, H. Wang, S. Xu, and W. Wang, “Interference avoidance mechanisms in

the hybrid cellular and device-to-device systems,” in Personal, Indoor and Mobile Radio

Communications, 2009 IEEE 20th International Symposium on, Sept 2009, pp. 617–621.

[21] C. F. M. Silva, José Mairton B. Silva Jr., and T. F. Maciel, “Radio resource management

for device-to-device communications in long term evolution networks,” in Resource

Allocation and MIMO for 4G and Beyond, F. R. P. Cavalcanti, Ed. New York: Springer

Science+Business Media, 2014, pp. 105–156.

[22] G. Fodor, E. Dahlman, G. Mildh, S. Parkvall, N. Reider, G. Mikló, and Z. Turányi, “Design

Aspects of Network Assisted Device-to-Device Communications,” IEEE Communications

Magazine, vol. 50, no. 3, pp. 170 –177, March 2012.

[23] K. Doppler, C.-H. Yu, C. Ribeiro, and P. Janis, “Mode selection for device-to-device

communication underlaying an lte-advanced network,” in Wireless Communications and

Networking Conference (WCNC), 2010 IEEE, April 2010, pp. 1–6.

[24] S. Hakola, T. Chen, J. Lehtomaki, and T. Koskela, “Device-to-device (d2d) communication

in cellular network - performance analysis of optimum and practical communication

Bibliography 55

mode selection,” in Wireless Communications and Networking Conference (WCNC), 2010

IEEE, April 2010, pp. 1–6.

[25] B. Wang, L. Chen, X. Chen, X. Zhang, and D. Yang, “Resource allocation optimization for

device-to-device communication underlaying cellular networks,” in Vehicular Technology

Conference (VTC Spring), 2011 IEEE 73rd, May 2011, pp. 1–6.

[26] A. Asadi and V. Mancuso, “A Survey on Opportunistic Scheduling in Wireless

Communications,” IEEE Communications Surveys Tutorials, vol. 15, no. 4, pp.

1671–1688, 2013.

[27] J. M. B. da Silva Jr., T. F. Maciel, R. L. Batista, C. F. M. e Silva, and F. R. P.

Cavalcanti, “UE grouping and mode selection for D2D communications underlaying a

multicellular wireless system,” in IEEE WCNC 2014 - Workshop on Device-to-Device and

Public Safety Communications (WCNC’14 - WDPC Workshop), Istambul, Turkey, Apr. 2014,

pp. 230–235.

[28] J. Gu, S. J. Bae, B.-G. Choi, and M. Y. Chung, “Dynamic power control mechanism

for interference coordination of device-to-device communication in cellular networks,” in

Ubiquitous and Future Networks (ICUFN), 2011 Third International Conference on, 2011,

pp. 71–75.

[29] N. Lee, X. Lin, J. G. Andrews, and R. W. H. Jr, “Power Control for D2D Underlaid Cellular

Networks: Modeling, Algorithms and Analysis,” http://arxiv.org/abs/1305.6161, may

2013.

[30] F. Baccelli, N. Khude, R. Laroia, J. Li, T. Richardson, S. Shakkottai, S. Tavildar, and

X. Wu, “On the design of device-to-device autonomous discovery,” in Communication

Systems and Networks (COMSNETS), 2012 Fourth International Conference on, Jan 2012,

pp. 1–9.

[31] M. Zulhasnine, C. Huang, and A. Srinivasan, “Efficient resource allocation for

device-to-device communication underlaying LTE network,” in Proceedings of the IEEE

Wireless and Mobile Computing, Networking and Communications (WiMob), Oct 2010, pp.

368–375.

[32] P. Janis, V. Koivunen, C. B. Ribeiro, K. Doppler, and K. Hugl, “Interference-avoiding

mimo schemes for device-to-device radio underlaying cellular networks,” in IEEE 20th

International Symposium on Personal, Indoor and Mobile Radio Communications, 2009,

Sep. 2009, pp. 2385 – 2389.

[33] B. Wang, L. Chen, X. Chen, X. Zhang, and D. Yan, “Resource Allocation Optimization

for Device-to-Device Communication Underlaying Cellular Networks,” in Vehicular

Technology Conference (VTC Spring), 2011 IEEE 73rd, May 2011, pp. 1–6.

[34] P. Jänis, C.-H. Yu, K. Doppler, C. Ribeiro, C. Wijting, K. Hugl, O. Tirkkonen, and

V. Koivunen, “Device-to-device communication underlaying cellular communications

systems,” in International Journal of Communications, Network and System

Sciences,2009, vol. 2, no. 3, Jun. 2009, pp. 169–178.

[35] S. Hakola, T. Chen, J. Lehtomaki, and T. Koskela, “Device-to-device (D2D)

communication in cellular network - performance analysis of optimum and practical

Bibliography 56

communication mode selection,” in Proceedings of the IEEE Wireless Communications and

Networking Conference (WCNC), Apr. 2010, pp. 1 –6.

[36] K. Doppler, C.-H. Yu, C. B. Ribeiro, and P. Janis, “Mode selection for device-to-device

communication underlaying an LTE-advanced network,” in Proceedings of the IEEE

Wireless Communications and Networking Conference (WCNC), Apr. 2010, pp. 1 –6.

[37] C.-H. Yu, K. Doppler, C. Ribeiro, and O. Tirkkonen, “Resource sharing

optimization for device-to-device communication underlaying cellular networks,”

Wireless Communications, IEEE Transactions on, vol. 10, no. 8, pp. 2752–2763, 2011.

[38] X. Xiao, X. Tao, and J. Lu, “A QoS-Aware Power Optimization Scheme in OFDMA Systems

with Integrated Device-to-Device (D2D) Communications,” in Vehicular Technology

Conference (VTC Fall), 2011 IEEE, 2011, pp. 1–5.

[39] C. Yu, O. Tirkkonen, K. Doppler, and C. Ribeiro, “On the Performance of Device-to-Device

Underlay Communication with Simple Power Control,” in Vehicular Technology

Conference, 2009. VTC Spring 2009. IEEE 69th, Apr. 2009, pp. 1–5.

[40] C. Li, B. Li, B. Lan, Y. Zhang, and T. Wang, “Uplink power control for device to device

communication underlaying cellular networks,” in Communications and Networking in

China (CHINACOM), 2013 8th International ICST Conference on, Aug 2013, pp. 256–259.

[41] G. Fodor, D. Della Penda, M. Belleschi, M. Johansson, and A. Abrardo, “A

comparative study of power control approaches for device-to-device communications,”

Communications (ICC), 2013 IEEE International Conference on, pp. 6008 – 6013, 2013.

[42] S. Gupta, R. Yates, and C. Rose, “Soft Dropping Power Control - A Power Control Backoff

Strategy,” in IEEE International Conference on Personal Wireless Communications, 1997,

Dec. 1997, pp. 210 –214.

[43] Y. V. L. Melo, R. L. Batista, T. F. Maciel, C. F. M. e Silva, J. M. B. da Silva Jr., and F. R. P.

Cavalcanti, “Power control with variable target SINR for D2D communications underlying

cellular networks,” in European Wireless 2014 (EW2014), Barcelona, Spain, May 2014.

[44] X. Lu, E. Kunnari, J. Leinonen, O. Piirainen, M. Vainikka, and M. Juntti, “Lte uplink

power control and base station antenna down tilt in a 3d channel model,” in Wireless

Conference (EW), 2010 European, April 2010, pp. 377–381.

[45] J. Salo, G. Del Galdo, J. Salmi, P. Kyösti, M. Milojevic, D. Laselva, and C. Schneider,

“MATLAB implementation of the 3GPP Spatial Channel Model (3GPP TR 25.996),”

On-line, Jan. 2005, http://www.tkk.fi/Units/Radio/scm/.

[46] B. Yu, L. Yang, H. Ishii, and X. Cheng, “Load balancing with antenna tilt control in

enhanced local area architecture,” in Vehicular Technology Conference (VTC Spring), 2014

IEEE 79th, May 2014, pp. 1–6.

[47] M. Garcia-Lozano, S. Ruiz, and J. Olmos, “UMTS optimum cell load balancing

for inhom*ogeneous traffic patterns,” in Vehicular Technology Conference, 2004.

VTC2004-Fall. 2004 IEEE 60th, vol. 2, Sept 2004, pp. 909–913 Vol. 2.

[48] K. S., G. C. D., and V. M. P., “Optimization by simulated annealing,” Science, 1983.

Bibliography 57

[49] M. Danneberg, J. Holfeld, M. Grieger, M. Amro, and G. Fettweis, “Field trial evaluation

of ue specific antenna downtilt in an lte downlink,” in Smart Antennas (WSA), 2012

International ITG Workshop on, March 2012, pp. 274–280.

[50] R. L. Batista, “Coordinated multipoint transmission and reception (comp) in a wireless

telecommunications network,” US Patent P37 998, Dec., 2012.

[51] A. A. Gebremariam, D. W. Kifle, B. Wegmann, I. Viering, and F. Granelli, “Techniques of

candidate cell selection for antenna tilt adaptation in lte-advanced,” in European Wireless

2014; 20th European Wireless Conference; Proceedings of, May 2014, pp. 1–6.

[52] I. Forkel, A. Kemper, R. Pabst, and R. Hermans, “The effect of electrical and mechanical

antenna down-tilting in umts networks,” in 3G Mobile Communication Technologies, 2002.

Third International Conference on (Conf. Publ. No. 489), May 2002, pp. 86–90.

[53] D. W. Kifle, B. Wegmann, I. Viering, and A. Klein, “On the potential of traffic driven

tilt optimization in lte-a networks,” in Personal Indoor and Mobile Radio Communications

(PIMRC), 2013 IEEE 24th International Symposium on, Sept 2013, pp. 2909–2913.

[54] ITU, “Framework and overall objectives of the future development of imt-2000 and

systems beyond IMT-2000, recommendation itu-r m.1645,” Tech. Rep., 2006.

[55] V. Hasu, “Radio resource management in wireless communication: Beamforming,

transmission power control, and rate allocation.” Ph.D. dissertation, Helsinki University

of Technology, 2007.

[56] J. Laiho and T. N. Achim Wacker, Radio Network Planning and Optimisation for UMTS.

John Wiley & Sons, 2006.

[57] R. C. David Soldani, Man Li, QoS and QoE Management in UMTS Cellular Systems. John

Wiley & Sons, 2006.

[58] 3GPP, “Physical layer aspect for evolved Universal Terrestrial Radio Access (UTRA),”

3rd Generation Partnership Project (3GPP), TR 25.814, Oct. 2006. [Online]. Available:

http://www.3gpp.org/ftp/Specs/html-info/25814.htm

[59] ——, “Spatial channel model for Multiple Input Multiple Output (MIMO) simulations,”

3rd Generation Partnership Project (3GPP), TR 25.996, Dec. 2009. [Online]. Available:

http://www.3gpp.org/ftp/Specs/html-info/25996.htm

[60] ——, “Further advancements for E-UTRA physical layer aspects,” Third Generation

Partnership Project, TR 36.814 V9.0.0, Mar. 2009.

[61] C. Mehlführer, M. Wrulich, J. C. Ikuno, D. Bosanska, and M. Rupp,

“Simulating the long term evolution physical layer,” in Procedings of the European

Signal Processing Conference, Glasgow, Scotland, Aug. 2009. [Online]. Available:

http://publik.tuwien.ac.at/files/PubDat_175708.pdf

[62] 3GPP, “Evolved Universal Terrestrial Radio Access (E-UTRA); Physical layer procedures,”

3rd Generation Partnership Project (3GPP), TS 36.213, Oct. 2010. [Online]. Available:

http://www.3gpp.org/ftp/Specs/html-info/36213.htm

[63] ——, “Evolved Universal Terrestrial Radio Access (E-UTRA); physical layer procedures,”

Third Generation Partnership Project, TS 36.213 V8.6.0, Mar. 2009.

Bibliography 58

[64] ——, “Physical layer aspects for evolved universal terrestrial radio access (UTRA),” Third

Generation Partnership Project, TR 25.814 V7.1.0, Sep. 2006.

[65] J. C. Ikuno, M. Wrulich, and M. Rupp, “System level simulation of LTE networks,” in

Proceedings of the IEEE Vehicular Technology Conference (VTC), May 2010, pp. 1–5.

[66] 3GPP, “Spatial channel model for multiple input multiple output (MIMO) simulations,”

Third Generation Partnership Project, TR 25.996 V7.0.0, Jun. 2007.

[67] G. Fodor, D. Della Penda, M. Belleschi, M. Johansson, and A. Abrardo, “A

comparative study of power control approaches for device-to-device communications,”

in Communications (ICC), 2013 IEEE International Conference on, June 2013, pp.

6008–6013.

[68] T. Maciel and A. Klein, “On the performance of SDMA with soft dropping and SINR

balancing power control in the downlink of multi-user MIMO systems,” in IEEE Workshop

on Smart Antennas (WSA 2007), Vienna, Austria, Feb. 2007.

[69] M. da S Rego, T. Maciel, H. de H M Barros, F. R. P. Cavalcanti, and G. Fodor, “Performance

analysis of power control for device-to-device communication in cellular mimo systems,”

in Wireless Communication Systems (ISWCS), 2012 International Symposium on, 2012,

pp. 336–340.

[70] C. F. Silva, J. M. B. Silva, and T. F. Maciel, Radio Resource Management

for Device-to-Device Communications in Long Term Evolution Networks, 1st ed.,

F. R. P. Cavalcanti, Ed. Springer New York, 2014. [Online]. Available:

http://dx.doi.org/10.1007/978-1-4614-8057-0_3

[71] W. Jianhui and Y. Dongfeng, “Antenna downtilt performance in urban environments,” in

Military Communications Conference, 1996. MILCOM ’96, Conference Proceedings, IEEE,

vol. 3, Oct 1996, pp. 739–744 vol.3.

[72] J. Niemela and J. Lempiainen, “Impact of mechanical antenna downtilt on performance

of wcdma cellular network,” in Vehicular Technology Conference, 2004. VTC 2004-Spring.

2004 IEEE 59th, vol. 4, May 2004, pp. 2091–2095 Vol.4.

[73] F. Gunnarsson, M. Johansson, A. Furuskar, M. Lundevall, A. Simonsson, C. Tidestav,

and M. Blomgren, “Downtilted base station antennas - a simulation model proposal and

impact on hspa and lte performance,” pp. 1–5, Sept 2008.

[74] 3GPP, “Evolved Universal Terrestrial Radio Access (e-utra); User Equipment (UE) Radio

Transmission and Reception,” Third Generation Partnership Project, Tech. Rep. TR

36.101 V12.2.0, Dec. 2013.

[75] D. Jaramillo-Ramirez, M. Kountouris, and E. Hardouin, “Coordinated multi-point

transmission with imperfect csi and other-cell interference,” Wireless Communications,

IEEE Transactions on, vol. 14, no. 4, pp. 1882–1896, April 2015.

[76] R. Fritzsche, P. Rost, and G. Fettweis, “Robust rate adaptation and proportional fair

scheduling with imperfect csi,” Wireless Communications, IEEE Transactions on, vol. PP,

no. 99, pp. 1–1, 2015.

Power Control and Energy Efficiency Strategies for D2D ... · Em outras palavras, a tecnologia D2D pode ser utilizada para aumentar ainda mais a eficiência de espectro e a eficiência - [PDF Document] (2024)

FAQs

What are examples of energy efficiency? ›

Energy efficiency, on the other hand, involves using technology that requires less energy to perform the same function. Energy-saving light bulbs, large household appliances, smart thermostats, and smart home hubs like Constellation Connect are all examples of technology that can be energy efficient.

How to be energy efficient? ›

Our top 10 energy saving tips
  1. To save the most energy, switch off the power point rather than leave appliances on standby.
  2. Turn off your heater, cooling units and appliances when you go to bed or leave the house.
  3. Switch off your computer and equipment such as printers or Wi-Fi routers overnight or when you're away.
Nov 17, 2023

What is efficiency in energy? ›

Energy efficiency means using less energy for the same output or producing more with the same energy input, and minimising energy waste. Reducing energy consumption and energy waste across the energy system — from production to final consumption — in all economic sectors is one of the EU's strategic objectives.

How does saving electricity improve human health? ›

The impact of energy efficiency in buildings

The potential benefits of energy efficiency measures include improved physical health such as reduced symptoms of respiratory and cardiovascular conditions, rheumatism, arthritis and allergies, as well as fewer injuries.

Can you have 100% energy efficiency? ›

Efficiencies cannot exceed 100%, which would result in a perpetual motion machine, which is impossible.

What kind of energy is 100% efficient? ›

Electric resistance heating is 100% energy efficient in the sense that all the incoming electric energy is converted to heat. However, most electricity is produced from coal, gas, or oil generators that convert only about 30% of the fuel's energy into electricity.

What wastes the most energy in a house? ›

What Can I Unplug? These Household Items Cost the Most Electricity
  • Cooling and heating: 47% of energy use.
  • Water heater: 14% of energy use.
  • Washer and dryer: 13% of energy use.
  • Lighting: 12% of energy use.
  • Refrigerator: 4% of energy use.
  • Electric oven: 3-4% of energy use.
  • TV, DVD, cable box: 3% of energy use.
Sep 1, 2022

What is the best energy saving device? ›

Types of energy-efficient products
  • Efficient light bulbs. ...
  • Advanced power strips (APS) ...
  • Smart switches. ...
  • Low-flow faucets and shower heads. ...
  • Smart thermostats. ...
  • Energy monitors. ...
  • Electric vehicles and chargers. ...
  • Solar energy systems.
Dec 6, 2023

How to reduce power consumption at home? ›

Tips for Saving on Your Electric Bill
  1. Turn Down Your Thermostat. It's one of the most effective ways to cut your energy usage. ...
  2. Take Care of Your Furnace. ...
  3. Keep the Cold Out. ...
  4. Turn Down the Tank. ...
  5. Cook Smart. ...
  6. Think Before You Wash and Dry Clothes. ...
  7. Reduce Phantom Load.

What is the useful power output? ›

useful output energy refers to the useful energy that is transferred by the device (eg thermal energy by a heater) input energy refers to the total energy supplied to a device.

How to calculate power wasted? ›

This can be done using the given efficiency and total input power. Calculate the wasteful power output. Simply subtract the useful power output from the total input power.

What is the most cost-effective light bulb? ›

LED light bulbs are the most energy-efficient option available, offering 40-80 lumens per watt and providing a long-term cost savings for homeowners.

How does electricity affect human? ›

Headaches, muscle fatigue or spasms, temporary unconsciousness, temporary breathing difficulty, severe burns, vision loss, hearing loss, brain damage, respiratory arrest or failure, cardiac arrest (heart attack), death.

What are the disadvantages of using energy efficient appliances? ›

Cons: Increased working times (for example, increased wash cycle time for dishwashers and washing machines) Their functioning changes to reduce energy usage, which can affect their performance. More expensive to buy.

Is clean energy actually clean? ›

Overall, clean energy is considered better for the environment than traditional fossil-fuel–based resources, generally resulting in less air and water pollution than combustible fuels, such as coal, natural gas, and petroleum oil.

What are the best example of efficiency? ›

In general, we say something is efficient when it maximises outputs with given inputs. In other words, it's the ability to do something well and without waste. Often we try to measure efficiency levels, such as how energy efficient our light bulbs are or how efficient a business is at producing a product.

What is an example of energy efficiency worked? ›

Example: An older piece of equipment receives 500 joules of power to produce the equivalent of 100 joules of output. 100/500 = 0.2, or 20% efficiency. A newer equipment version takes the same 500-joule input to generate 400 joules of productive output. 400/500 = 0.8, or 80% efficiency—much better!

What is an example of an energy efficient device? ›

Clothes Dryers

Energy-efficient dryers use less energy than conventional models without sacrificing features or performance. They do this by deploying technologies such as moisture sensors that detect when clothes are dry and automatically shut the dryer off.

What is simple energy efficiency? ›

What can I do? Turn off the power switch on the socket/wall or unplug appliances from the socket when they are not in use. Turning devices to “energy-saving” or “eco” mode is another way to reduce energy consumption.

Top Articles
Latest Posts
Article information

Author: Eusebia Nader

Last Updated:

Views: 5919

Rating: 5 / 5 (60 voted)

Reviews: 91% of readers found this page helpful

Author information

Name: Eusebia Nader

Birthday: 1994-11-11

Address: Apt. 721 977 Ebert Meadows, Jereville, GA 73618-6603

Phone: +2316203969400

Job: International Farming Consultant

Hobby: Reading, Photography, Shooting, Singing, Magic, Kayaking, Mushroom hunting

Introduction: My name is Eusebia Nader, I am a encouraging, brainy, lively, nice, famous, healthy, clever person who loves writing and wants to share my knowledge and understanding with you.