# Nguyen, Son, Capacity and Throughput Optimization in Multi-cell 3G WCDMA Networks

**Master of Science (Computer Science and Engineering), August 2005, 82 pp., 16 tables, 25
**

ﬁgures, 50 titles.

User modeling aids in the computation of the traﬃc density in a cellular network, which

can be used to optimize the placement of base stations and radio network controllers as well

as to analyze the performance of resource management algorithms towards meeting the ﬁnal

goal: the calculation and maximization of network capacity and throughput for diﬀerent

data rate services. An analytical model is presented for approximating the user distributions

in multi-cell third generation Wideband Code Division Multiple Access (WCDMA) networks

using 2-dimensional Gaussian distributions by determining the means and the standard de-

viations of the distributions for every cell. This allows for the calculation of the inter-cell

interference and the reverse-link capacity of the network.

An analytical model for optimizing capacity in multi-cell WCDMA networks is presented.

Capacity is optimized for diﬀerent spreading factors and for perfect and imperfect power

control. Numerical results show that the SIR threshold for the received signals is decreased by

0.5 to 1.5 dB due to the imperfect power control. The results also show that the determined

parameters of the 2-dimensional Gaussian model matches well with traditional methods for

modeling user distribution.

A Call Admission Control algorithm is designed that maximizes the throughput in multi-

cell WCDMA networks. Numerical results are presented for diﬀerent spreading factors and

for several mobility scenarios. Our methods of optimizing capacity and throughput are

computational eﬃcient, accurate, and can be implemented in large WCDMA networks.

CAPACITY AND THROUGHPUT OPTIMIZATION IN MULTI-CELL

3G WCDMA NETWORKS

Son Nguyen, B.S.

Thesis Prepared for the Degree of

MASTER OF SCIENCE

UNIVERSITY OF NORTH TEXAS

August 2005

APPROVED:

Robert Akl, Major Professor

Robert Brazile, Committee Member and

Graduate Coordinator

Steve Tate, Committee Member

Krishna Kavi, Chair of the Department

of Computer Sciences

Oscar N. Garcia, Dean of the College of

Engineering

Sandra L. Terrell, Dean of the Robert B. Toulouse

School of Graduate Studies

ACKNOWLEDGMENTS

First and foremost, I would like to give my heartfelt thanks to my advisor Dr. Robert Akl

for taking on multitude of roles that provided guidance and direction. During this journey,

I had sometimes felt I could not progress any further, but his whole-hearted devotion and

enthusiasm not only kept me on track but also lightened up the way to the completion of

this work. Furthermore, I am highly indebted to the members of my committee Dr. Robert

Brazile and Dr. Steve Tate for their careful reading and suggestions.

I am also very grateful to my parents for their unconditional love and many years of

support. Above all, I would like to thank Khanh Ha Nguyen who has been an inspiration

and a partner from the very beginning.

ii

CONTENTS

ACKNOWLEDGEMENTS iii

LIST OF TABLES vii

LIST OF FIGURES ix

1 INTRODUCTION 1

1.1 CDMA History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.2 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.3 Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

2 CDMA AND WCDMA OVERVIEW 4

2.1 Introduction to CDMA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2.1.1 Power Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.1.2 Frequency Reuse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

2.1.3 Voice Activity Detection . . . . . . . . . . . . . . . . . . . . . . . . . 8

2.1.4 Cell Sectoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2.1.5 Soft Handoﬀ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2.2 WCDMA Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

3 USER AND INTERFERENCE MODELING USING 2-D GAUSSIAN FUNCTION 17

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

3.2 Related Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

3.3 User and Interference Model . . . . . . . . . . . . . . . . . . . . . . . . . . 18

3.4 Numerical Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

3.4.1 Uniform Distribution of Users . . . . . . . . . . . . . . . . . . . . . . 21

iii

3.4.2 Users Densely Clustered at the Center of the Cells . . . . . . . . . . 21

3.4.3 Users Distributed at Cells’ Boundaries . . . . . . . . . . . . . . . . . 26

3.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

4 WCDMA CAPACITY 29

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

4.2 Related Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

4.3 WCDMA Capacity with Perfect Power Control . . . . . . . . . . . . . . . . 31

4.4 WCDMA Capacity with Imperfect Power Control . . . . . . . . . . . . . . . 32

4.5 Spreading and Scrambling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

4.6 Numerical Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

4.6.1 WCDMA Capacity Optimization with SF of 256 . . . . . . . . . . . . 37

4.6.2 WCDMA Capacity Optimization with SF of 64 . . . . . . . . . . . . 37

4.6.3 WCDMA Capacity Optimization with SF of 16 . . . . . . . . . . . . 39

4.6.4 WCDMA Capacity Optimization with SF of 4 . . . . . . . . . . . . . 41

4.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

5 WCDMA CALL ADMISSION CONTROL AND THROUGHPUT 46

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

5.2 Feasible States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

5.3 Mobility Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

5.4 WCDMA Call Admission Control . . . . . . . . . . . . . . . . . . . . . . . . 49

5.5 Network Throughput . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

5.6 Calculation of N . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

5.7 Maximization of Throughput . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

5.8 Numerical Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

5.8.1 WCDMA Throughput Optimization with SF of 256 . . . . . . . . . . 54

iv

5.8.2 WCDMA Throughput Optimization with SF of 64 . . . . . . . . . . . 54

5.8.3 WCDMA Throughput Optimization with SF of 16 . . . . . . . . . . . 56

5.8.4 WCDMA Throughput Optimization with SF of 4 . . . . . . . . . . . 56

5.9 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

6 CONCLUSIONS 62

6.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

6.2 Future Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

v

LIST OF TABLES

2.1 Main diﬀerences between WCDMA and IS-95 air interfaces . . . . . . . . . . 13

3.1 The maximum number of users in every cell for the 27 cell WCDMA network

(with σ

1

and σ

2

are increased from 5000 to 15000 while µ

1

= 0 and µ

2

= 0).

This results in users distributed uniformly in all BSs. . . . . . . . . . . . . . 22

3.2 The maximum number of users in 27 cells of WCDMA network as the values

of σ

1

and σ

2

are increased from 100 to 400 while µ

1

= 0 and µ

2

= 0. This

results in users densely clustered around the BSs. . . . . . . . . . . . . . . . 24

3.3 The values of σ

1

, σ

2

, µ

1

, and µ

2

for the 2-D Gaussian approximation of users

clustered at the boundaries of the cells as shown in Fig. 3.6. The maximum

number of users is 133. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

4.1 Functionality of the channelization and scrambling codes. . . . . . . . . . . . 34

4.2 Uplink DPDCH data rates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

4.3 Capacity calculation for uniform user distribution with SF = 256 and

E

b

Io

=

7.5 dB. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

4.4 Capacity calculation for uniform user distribution with SF = 64 and

E

b

Io

= 7.5

dB. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

4.5 Capacity calculation for uniform user distribution with SF = 16 and

E

b

Io

= 7.5

dB. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

4.6 Capacity calculation for uniform user distribution with SF = 4 and

E

b

Io

= 7.5

dB. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

5.1 The low mobility characteristics and parameters. . . . . . . . . . . . . . . . . 53

5.2 The high mobility characteristics and parameters. . . . . . . . . . . . . . . . 54

vi

5.3 Calculation of N for uniform user distribution with SF = 256 and blocking

probability = 0.02. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

5.4 Calculation of N for uniform user distribution with SF = 64 and blocking

probability = 0.02. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

5.5 Calculation of N for uniform user distribution with SF = 16 and blocking

probability = 0.02. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

5.6 Calculation of N for uniform user distribution with SF = 4 and blocking

probability = 0.02. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

vii

LIST OF FIGURES

2.1 Comparison between FDMA, TDMA, and CDMA. . . . . . . . . . . . . . . 4

2.2 Frequency Hopping Spreading Spectrum. . . . . . . . . . . . . . . . . . . . . 6

2.3 Time Hopping Spreading Spectrum. . . . . . . . . . . . . . . . . . . . . . . . 6

2.4 2 GHz band spectrum allocation in Europe, Japan, Korea, and US (MSS =

Mobile Satellite Spectrum). . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

2.5 Development to all-IP for 3G services. . . . . . . . . . . . . . . . . . . . . . 15

3.1 Inter-cell interference on cell i from users in cell j . . . . . . . . . . . . . . . 19

3.2 2-D Gaussian approximation of users uniformly distributed in the cells. σ

1

=

σ

2

= 12000, µ

1

= µ

2

= 0. The maximum number of users is 548. . . . . . . . 23

3.3 Simulated network capacity where users are uniformly distributed in the cells.

The maximum number of users is 554. . . . . . . . . . . . . . . . . . . . . . 23

3.4 2-D Gaussian approximation of users densely clustered around the BSs. σ

1

=

σ

2

= 100, µ

1

= µ

2

= 0. The maximum number of users is 1026. . . . . . . . 25

3.5 Simulated network capacity where users are densely clustered around the BSs

causing the least amount of inter-cell interference. The maximum number of

users is 1026 in the network. . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

3.6 2-D Gaussian approximation of users clustered at the boundaries of the cells.

The values of σ

1

, σ

2

, µ

1

, and µ

2

may be diﬀerent in the diﬀerent cells and are

given in Table 3.3. The maximum number of users is 133. . . . . . . . . . . . 26

3.7 Simulated network capacity where users are clustered at the boundaries of

the cells causing the most amount of inter-cell interference. The maximum

number of users is only 108 in the network. . . . . . . . . . . . . . . . . . . . 28

4.1 Generation of OVSF codes for diﬀerent Spreading Factors. . . . . . . . . . . 33

4.2 Relationship between spreading and scrambling. . . . . . . . . . . . . . . . . 34

viii

4.3 12.2 Kbps Uplink Reference channel. . . . . . . . . . . . . . . . . . . . . . . 35

4.4 64 Kbps Uplink Reference channel. . . . . . . . . . . . . . . . . . . . . . . . 36

4.5 Average number of slot per sector for perfect and imperfect power control

analysis with a Spreading Factor of 256. . . . . . . . . . . . . . . . . . . . . 39

4.6 Average number of slot per sector for perfect and imperfect power control

analysis with a Spreading Factor of 64. . . . . . . . . . . . . . . . . . . . . . 41

4.7 Average number of slot per sector for perfect and imperfect power control

analysis with a Spreading Factor of 16. . . . . . . . . . . . . . . . . . . . . . 43

4.8 Average number of slot per sector for perfect and imperfect power control

analysis with a Spreading Factor of 4. . . . . . . . . . . . . . . . . . . . . . . 45

5.1 Average throughput in each cell for SF = 256. . . . . . . . . . . . . . . . . . 56

5.2 Average throughput in each cell for SF = 64. . . . . . . . . . . . . . . . . . . 57

5.3 Average throughput in each cell for SF = 16. . . . . . . . . . . . . . . . . . . 57

5.4 Average throughput in each cell for SF = 4. . . . . . . . . . . . . . . . . . . 60

ix

CHAPTER 1

INTRODUCTION

1.1 CDMA History

The global mobile communications market has expanded very rapidly [13]. From analog

phone systems in the 70’s and 80’s, cellular phone systems have progressed to digital cellu-

lar systems in their second generation (2G) with Time Division Multiple Access (TDMA),

Frequency Division Multiple Access (FDMA), and Code Division Multiple Access (CDMA)

technologies in the 90’s. Society has seen the introduction of data services from 2.5G with

Short Message Service and now new third generation (3G) mobile phone systems are being

introduced where cellular phones can access Internet services, retrieving text, pictures, video

and other documents, with a promised delivery speed of up to 2 Mbps.

Almost 1.52 billion people were using mobile phones by the end of 2004, and according

to the new study by the Yankee Group, this number could reach 1.87 billion (27.4% total

world population) by the end of 2007 [30]. CDMA is the fastest-growing digital wireless

technology since its ﬁrst commercialization in 1994. 240 million subscribers already existed

worldwide by the end of 2004 [15]. The major markets for CDMA are North America, Latin

America, and Asia (particularly Japan and Korea). In total, CDMA was adopted by more

than 100 operators across 76 countries around the globe [16]. According to [25, 50], CDMA

technology can oﬀer about 7 to 10 times the capacity of analog technologies and up to 6

times the capacity of digital technologies such as TDMA. The advantages over TDMA and

FDMA technologies, such as voice quality, system reliability, and handset battery life, have

created a multitude of research on CDMA systems.

1

1.2 Objectives

In this work, the 2-dimensional (2-D) Gaussian function is used to model user distribution

in cellular networks. In addition, capacity in Wideband Code Division Multiple Access

(WCDMA) cellular networks is calculated and optimized for diﬀerent data rate services.

The objectives of this work are as follows:

• User distribution modeling with 2-D Gaussian function:

– Modeling user distribution by determining the user densities, the means, and the

standard deviations of the 2-D Gaussian function in every cell.

• WCDMA capacity:

– Formulation and calculation of the maximum capacity for diﬀerent services in

WCDMA networks based on given quality of service (QoS) constraints.

• WCDMA call admission control and throughput:

– Formulation and calculation of Call Admission Control (CAC) in WCDMA net-

works based on the QoS and Grade of Service (GoS) constraints. This leads to

the maximization of the throughput in WCDMA networks.

1.3 Organization

In Chapter 2, CDMA and WCDMA technologies are introduced. While CDMA is a 2G

cellular system with a main focus on voice services, WCDMA is the emerging technology

for 3G mobile phone systems with diﬀerent data rates on demand for serving diﬀerent ser-

vices for users. This chapter summarizes features as well as key points, which demonstrate

CDMA’s superiority over other 2G cellular networks. These superior features include power

control, user activity detection, soft handoﬀ, and cell sectoring. In addition, this chapter

2

also introduces the requirements for 3G cellular systems, which include wider bandwidth

allocations, as well as new additional features of 3G WCDMA cellular networks.

In Chapter 3, the 2-D Gaussian function is used in modeling users in cellular networks.

This method when used to calculate the average interference can more rapidly and eﬃciently

compute the capacity in WCDMA networks when compared to other methods, like actual

interference, which are computationally intensive.

In Chapter 4, the formulation and calculation of the maximum capacity in WCDMA

cellular networks is described. Our optimization with given quality of service constraints

can ﬁnd the maximum number of simultaneous users for voice and data services in 3G

networks.

In Chapter 5, a call admission control algorithm is designed that maximizes the through-

put in multi-cell WCDMA networks. Numerical results are presented for diﬀerent spreading

factors and for several mobility scenarios.

Finally, in Chapter 6, the conclusions are presented, which summarize the contributions

of this work.

3

CHAPTER 2

CDMA AND WCDMA OVERVIEW

2.1 Introduction to CDMA

Code Division Multiple Access (CDMA) is a fairly new wireless communication technology,

which was introduced by Qualcommin the early 90’s. Compared to existing technologies such

as FDMA and TDMA, which use diﬀerent frequency sub-bands and time slots respectively, to

carry multiple calls, CDMA distinguishes diﬀerent calls by unique codes, which are assigned

individually to each call. Fig. 2.1 depicts the diﬀerences between the three technologies.

CDMA has many advantages over TDMA and FDMA technologies in cellular networks.

These advantages include increased capacity, immunity to multi-path fading, voice activity,

and soft handoﬀ mechanism. These gains and features yield better longevity of CDMA

handset battery life and higher quality of voice signals. According to [25], an analysis

made by the Telecommunications Research and Action Center found CDMA outperformed

other digital and analog technologies on every front, including signal quality, security, power

consumption, and reliability.

Figure 2.1: Comparison between FDMA, TDMA, and CDMA.

4

CDMA is based on Spread Spectrum (SS) communication, a technique which was devel-

oped during World War II. The essential idea behind SS is spreading the information signal

over a wider bandwidth to make jamming and interception more diﬃcult. There are three

types of SS: Frequency Hopping, Time Hopping, and Direct Sequence (DS). The ﬁrst type

of spread spectrum is known as Frequency Hopping in which the signal is broadcasted over

a seemingly random series of radio frequencies, hopping from frequency to frequency at a

ﬁxed interval, as shown in Fig. 2.2. By knowing the hopping sequence, which is contained

in the spreading code, the receiver, hopping between frequencies in synchronization with

the transmitter, picks up the message. Time Hopping is the second type of SS where the

transmission time is divided into equal amounts of time intervals called frames. Each frame

is divided into time slots. During each frame, only one time slot is used to transmit the mes-

sage, as shown in Fig. 2.3. Time Hopped systems assume that the sender and the receiver

know the length of each time frame and the sequence of time slots in which the message will

be modulated. The third spreading technique is Direct Sequence. Direct Sequence Spread

Spectrum modulates the data with a fast pseudo-random code sequence. With DS, each bit

in the original signal is modulated to multiple bits in the transmitted signal by using the

spreading code. With the spreading code, the signal is spread across a wider frequency band

in direct proportion to the number of bits used. At this time, DS is the prefered spreading

technique in CDMA and WCDMA and thus is the only one considered in this work.

With DS-CDMA, each user’s signal is modulated with a high-rate special code (pseudo-

random binary sequence), which is assigned by the Base Station (BS), into a wideband signal

transmitted over the air medium. CDMA users share the same frequency spectrum; thus,

CDMA signals appear to overlap in the time and frequency domain as was shown in Fig.

2.1. Each user’s signal is distinguished by its special code. The only way to despread the

signal to recover the message for a user is to use each user’s own special code. Only the

subscribed BS knows all its users’ special codes. For each individual user, other signals from

5

Figure 2.2: Frequency Hopping Spreading Spectrum.

Figure 2.3: Time Hopping Spreading Spectrum.

6

diﬀerent users appear as noise and are represented as interference generated by the system.

In a CDMA network, the signal for each user is required to be above a given signal-to-

interference ratio (SIR) threshold for it to be received and despread by the BS correctly.

Therefore, the capacity of a CDMA network is limited by the amount of interference that is

generated by all users in the network (unlike FDMA and TDMA capacities, which are ﬁxed

and primarily bandwidth limited).

2.1.1 Power Control

Since the capacity of a CDMA network is interference-bound, the study of capacity charac-

teristics focuses primarily on the methods of reducing interference. Fast and precise power

control is a key requirement for CDMA technology. Power control aims to reduce interfer-

ence by minimizing the eﬀects of the near-far problem (the received power at a BS from

a mobile station (MS) near the cell boundary is less compared to a MS close to the BS),

co-channel interference, and fading while keeping the received signal power, or the SIR, at

the same level at the BS. In addition, power consumption can be signiﬁcantly reduced in the

MS, which results in increasing longevity of the MS. Research on power control algorithms

[19, 23, 27, 28, 52, 61, 77] has substantially investigated the near-far problem. In [54], the

authors propose an adaptive SIR based feedback power control, which tries to solve the

near-far problem while maintaining a low co-interference eﬀect by individually adjusting the

SIR threshold control level for each mobile station, with respect to its own radio link. This

control level is established by using Fuzzy Logic Control. In [8], the authors investigate a

feedback power control approach that allows power commands to be updated at a faster rate

compared to the rate of multipath fading. In [73, 76] the authors deﬁne the problem of SIR

balancing to be an eigenvalue problem in a link-gain matrix, and try to ﬁnd the optimum

downlink power control.

There are two main diﬀerent methods for managing power control: open-loop and closed-

7

loop. Open-loop power control determines the transmit power such that the sum of transmit

power and the received power is kept constant. Closed-loop power control involves with two

parties - BS and MS, in which one party signiﬁes the other party to increase or decrease its

transmit power by a power step such that the target received

E

b

Io

achieves a given threshold.

2.1.2 Frequency Reuse

In TDMA and FDMA, the total bandwidth is divided into a number of channels. Each user

is assigned a channel for an uplink (the frequency that a MS sends information to a BS)

and a downlink (the frequency that a BS transmits the signal to a MS). The transmitted

signal is attenuated with both fast (Rayleigh) and slow fading. Thus, with precise power

control, diﬀerent cells in a cellular network can use the same frequencies if the signals from

the same frequencies from one cell will not reach the BS of another cell. This concept is

called frequency reuse capability. In a hexagonal cellular network structure, the frequency

reuse factor may be 4, 7, or 12. Because a cellular network is given a pre-determined range of

frequencies to exchange information between the BSs and MSs, if the frequency reuse factor,

which the cellular network can employ, is smaller, then the network can obtain a higher

capacity. CDMA cellular systems typically use universal frequency reuse (or a frequency

reuse factor of 1), where the MSs and BSs use the whole bandwidth to transfer and receive

information.

2.1.3 Voice Activity Detection

In CDMA systems, reducing multiple access interference from neighboring cells results in a

capacity gain. Since CDMA systems use speech coding, reducing the rate of the speech coder

with voice activity detection along with variable rate data transmission could decrease the

multiple access interference. According to [33], the voice activity factor for human speech

averages about 42%. In [24, 33], the authors show that increasing the voice activity factor or

8

data activity factor could reduce the network capacity signiﬁcantly. In FDMA and TDMA

cellular systems, the frequencies are permanently assigned to users and BSs as long as there

is a communication between them. Thus, the capacity in FDMA and TDMA systems is

ﬁxed regardless whether the systems employ voice or data activity detection.

2.1.4 Cell Sectoring

Due to increasing demand for cellular communication without corresponding increases in

bandwidth allocation, the authors in [11, 12, 57] introduce Cell Sectoring (Spatial Processing)

to improve spectrum utilization. Cell Sectoring is a method that uses multiple directional

antenna arrays to reduce the co-channel interference, resulting in increased cellular network

capacity. According to [42, 56], resolving angular positions of the mobiles by using antenna

arrays at BSs, both in receiving and transmitting, can lead to many-fold increases in system

capacity.

2.1.5 Soft Handoﬀ

Soft handoﬀ (handover) is one of the most attractive features of CDMA technology. Soft

handoﬀ is a technique that allows a MS in transition between two or more adjacent cells to

transmit and receive the same signal from these BSs simultaneously. By employing universal

frequency reuse and a Rake receiver (see [58] pages 49-56), each individual MS can isolate

and align both in time and in phase to reinforce the forward signals from diﬀerent BSs, as

well as to transmit its signal to all the BSs. On the uplink side, the Mobile Switching Center

must ﬁrst combine and resolve the signals from all the BSs and then determine which BS is

receiving the stronger and better replica. Depending on the relative received signal strength,

decisions are made as to when to enter soft handoﬀ and when to release the weaker BS. Recent

improvements in soft handoﬀ algorithms are described in [9, 75]. Experimental data in [70]

have shown that using soft handoﬀ increases cell coverage resulting in increased capacity

9

in CDMA networks. In [17], a comparison between hard handoﬀ (in TDMA/FDMA) and

soft handoﬀ (in CDMA) shows that under a variety of conditions, the shadow fading margin

required by CDMA soft handoﬀ is less than TDMA/FDMA systems by 2.6-3.6 dB, which

translates to range extension for CDMA cellular networks. In [35], the authors compare the

network capacity gain with diﬀerent parameter sets for new soft handoﬀ algorithm in IS-95A

[64] and IS-95B [63]. Soft handoﬀ, however, requires complex design and implementation.

Recent surveys [39, 74] of soft handoﬀ show what technical issues need to be resolved. In

addition, the surveys show what the beneﬁts and tradeoﬀs of using soft handoﬀ and discuss

feasible parameter settings.

Extensive research has been done on calculating the reverse link capacity of single and

multi-cell CDMA networks [10, 20, 31, 46, 51, 53, 59, 60, 71, 72]. These studies conclude that

the capacity of the reverse link is lower than the capacity of the forward link. In [26], the

authors analyze both reverse link and forward link capacity under the assumption of ideal

power control and hard handoﬀ circumstances. The reverse link capacity limits the system

capacity; however, only a small diﬀerence exists between forward and reverse link capacity.

In this work, only the reverse link capacity is considered.

Most of the models [7, 38, 40, 62] use average inter-cell interference instead of actual

inter-cell interference for capacity analysis. In an average inter-cell interference model, the

interference caused by diﬀerent users in the same cell is identical, and is independent of their

exact location within a cell. Thus, accurate user modeling becomes essential for average

interference calculation. In this work, average inter-cell interference is used to model and

calculate capacity in WCDMA cellular networks.

2.2 WCDMA Overview

Third generation (3G) systems are designed for multimedia communication, which includes

person-to-person communication with high-quality images and video and high rate accessing

10

of information and services on public and private networks, up to 2 Mbps. These new

systems will create new business opportunities, not only for the cellphone, infrastructure,

and hardware manufactures and their operators, but also for the applications and content

providers carried by these networks (i.e., paid video games, music, video, and ring tones

downloading).

WCDMA speciﬁcations have been created in 3GPP (3rd Generation Partnership Project),

which is the joint standardization project of the bodies from Europe, Japan, Korea, the US,

and China. Within 3GPP, WCDMA is identiﬁed as UTRA (Universal Terrestrial Radio

Access) FDD (Frequency Division Duplex) and TDD (Time Division Duplex), the name

WCDMA is used to cover both FDD and TDD operations.

The process of developing third generation mobile systems started in a 1992 meeting

of the World Administrative Radio Conference (WARC), which is under the International

Telecommunication Union (ITU). In the meeting, the conference identiﬁed that the frequen-

cies around 2 GHz were available in most countries to be used by the future third generation

mobile systems, both terrestrial and satellite as shown in Fig. 2.4 [29]. Under ITU, these

third generation systems are called International Mobile Telephony 2000 (IMT-2000).

There are three major air interfaces proposed for 3G cellular systems: Enhanced Data

Rate for GSM Evolution (EDGE), W-CDMA, and CDMA2000 [34]. The spectrum allocation

for IMT-2000 (or WARC-92) bands of 2x60MHz (1920-1980 MHz plus 2110-2170 MHz) are

available in Europe, Japan, Korea, and most Asian countries. The two spectrums for IMT-

2000 TDD bands (1900-1920 MHz and 2020-2025 MHz) are available in Europe and Korea.

In Japan, part of the IMT-2000 spectrum TDD is used by cordless telephone systems. In

the United States, at the time of the 1992 meeting, no new spectrum had yet been made

available for third generation systems. In the United States, third generation services can

be implemented using the existing PCS spectrum with alternative technologies, including

EDGE, WCDMA, and CDMA2000.

11

Figure 2.4: 2 GHz band spectrum allocation in Europe, Japan, Korea, and US (MSS =

Mobile Satellite Spectrum).

At the ITU-R WRC-2000 in May 2000, the following additional frequency bands were

also introduced for IMT-2000: 1710-1885 MHz, 2500-2690 MHz, and 806-960 MHz

According to [29], new third generation spectrums in the United States are expected to

have 2x60 MHz (1710-1770 MHz and 2110-2170 MHz) assigned. These two spectrums can be

eﬃciently used to carry 3rd generation services with WCDMA. The new IMT-2000 spectrum

190 MHz (2500-2690 MHz) arrangement is still under discussion.

In the past, 2G cellular mobile phone systems mainly concentrated on voice traﬃc; now,

3G systems face the challenge of making data services wireless. In addition, the following

characteristics are the new requirements of 3G systems:

• Data communication speeds up to 2 Mbps.

• Variable bit data rates on demand.

• Mix of services with diﬀerent quality requirements on a single connection, including

voice, video, and packet data.

12

Table 2.1: Main diﬀerences between WCDMA and IS-95 air interfaces

WCDMA IS-95

Carrier Spacing 5 MHz 1.25 MHz

Chip rate 3.84 Mcps 1.2288 Mcps

Power control frequency 1500 Hz, both uplink and downlink Uplink: 800 Hz, downlink: slow power control

Base station synchronization Not needed Yes, typically obtained via GPS

Inter-frequency handovers Yes, measurements with slotted mode Possible, but measurement method not speciﬁed

Eﬃcient radio resource man-

agement algorithms

Yes, provides required quality of service Not needed for speech only networks

Packet data Load-based packet scheduling Packet data transmitted as short circuit switched

calls

Downlink transmit diversity Supported for improving downlink capacity Not supported by the standard

• Meet delay requirement constraints for diﬀerent real-time services, including delay-

sensitive real-time traﬃc (stock information) to best-eﬀort packet data (video or voice).

• Quality of service (QoS) requirements from 10% frame error rate to 10

−6

bit error rate.

• Coexistence with second generation systems and feasible solutions for inter-system

handovers, which include coverage enhancements and load balancing.

• Support asymmetric uplink and downlink traﬃc, e.g., web browsing results in more

downlink than uplink traﬃc.

• Higher spectrum eﬃciency.

• Coexistence of FDD and TDD modes.

The main diﬀerences between WCDMA and IS-95 are shown in Table 2.1 [29]. IS-95

is a second generation cellular network, which operates on the same frequency band as the

Advanced Mobile Phone System (ﬁrst generation cellular network), using FDD with a total

13

bandwidth of 25 MHz in each direction [25]. Both WCDMA and IS-95 use direct sequence

CDMA. WCDMA uses a higher chip rate of 3.84 Mcps. According to [29], the higher

chip rate in WCDMA enables higher bit rates while providing more multipath diversity,

which improves the coverage of a network when cellular equipments are employed with Rake

receivers [17].

WCDMA uses fast closed-loop power control in both uplink and downlink, while IS-95

uses fast power control only in the uplink. The inclusion of fast power control at the MS

side improves link performance and increases downlink capacity. However, it increases the

complexity in the design of the MS.

The IS-95 system was designed to target mainly macro cellular applications. The macro

cell BSs are required to be located on masts or rooftops where GPS signals can be eas-

ily received. IS-95 BSs are required to be synchronized and this can be done through the

GPS system [47]. Since GPS reception is diﬃcult without line-of-sight connection to the

GPS satellites, the deployment of indoor and micro cells was a challenge until recent tech-

nology, indirect GPS, was developed [47]. BSs in WCDMA are designed to operate under

asynchronous mode; thus, no synchronization from GPS is needed.

WCDMA changed the way communication is done in the core networks as shown in Fig.

2.5 [29]. In 2G and 2.5G CDMA cellular systems, all services (voice, SMS, WAP, and email)

were using circuit switched core networks. In March 2000, 3GPP Release ’99 kept voice and

video services under circuit switched, while transforming SMS, WAP, and email to packet

switched core networks. Also, Web, MMS, and Streaming services were added to packet

switched core networks. The latest releases 5 and 6 of 3GPP have made all services use

packet switched core networks [1, 29]. These releases have led to WCDMA implementations

to work with asynchronous BSs, where no synchronization from GPS is necessary. The

asynchronous BSs have made the handoﬀ in WCDMA diﬀerent from IS-95.

Inter-frequency handoﬀ is essential in WCDMA to maximize the use of several carriers

14

Figure 2.5: Development to all-IP for 3G services.

15

per BS, while in IS-95 inter-frequency measurements are not speciﬁed.

16

CHAPTER 3

USER AND INTERFERENCE MODELING USING 2-D GAUSSIAN FUNCTION

3.1 Introduction

Radio Network Planning is the problem of dimensioning, which is a process of ﬁnding possible

conﬁgurations and the amount of network equipment needed for an operator’s coverage,

capacity, quality of service, area type, and radio propagation requirements. Speciﬁcally,

dimensioning activities involve radio link budget and coverage analysis, capacity calculation,

and an estimation on the amount of sites, base station hardware, radio network controllers,

equipments at diﬀerent interfaces, and core network elements, including circuit switched

domain and packet switched (IP based) domain core network.

User Modeling in WCDMA is a major process in the dimensioning activities. With a

given network setting (i.e., available spectrum, blocking probability, and BS characteristics)

and a path loss model, the dimensioning activities determine the user distribution, which

is essential for the accurate calculation of interference and mobility. User Modeling helps

compute the traﬃc density in the cellular network, which can be used to optimize the place-

ment of BSs and radio network controllers as well as to analyze the performance of resource

management algorithms towards meeting the ﬁnal goal: the calculation and maximization

of network capacity.

3.2 Related Work

In the past decade, the problem of modeling user distribution and mobility has been an

engaging research subject.

17

In [18], the authors propose a method for creating two layers of hierarchical cellular

networks: Macrocell and Microcell to address high and low mobility of users. Microcells

accommodate both new and handoﬀ requests and can handle all types of high-rate connection

requests, including voice, video, and data, while Macrocells, which have larger coverage, can

be responsible for both new and handoﬀ requests but have limited application capabilities,

such as reduced quality video. This speciﬁc arrangement made the Macrocells encompass

and lower the handoﬀ rate for faster-moving users, while the Microcells can be deployed in

hot-spot areas to provide high-rate services to stationary or slow-moving users.

In [22], the authors use dynamic pricing to regulate the demand on wireless services, as

well as altering the mobility of users during peak time while also utilizing network resources

during oﬀ-peak hours by lowering prices to the users.

In [44], the authors’ model diﬀerent aspects of user distribution and mobility by taking

two new concepts into account: user classes and street types. With the ability to load real

map data into numerical simulations, the software can calculate the network capacity and

capture the movement of diﬀerent user classes in Helsinki area.

Recent research has looked at diﬀerent aspects that contribute to user distribution and

mobility, but most of it has focused on mobility and manually placing users in simulations.

In [6], the authors show that user distribution does drastically aﬀect the overall network

capacity. This section will show that by using the 2D-Gaussian function at each BS, user

distribution and interference can be easily modeled for many diﬀerent scenarios, including

users uniformly distributed, users densely clustered at the center of the cells, and users at

the cells’ boundaries.

3.3 User and Interference Model

This study assumes that each user is always communicating and is power controlled by the

base station (BS) that has the highest received power at the user. Let r

i

(x, y) and r

j

(x, y)

18

Figure 3.1: Inter-cell interference on cell i from users in cell j

be the distance from a user to BS i and BS j, respectively. This user is power controlled

by BS j in the cell or region C

j

with area A

j

, which BS j services as showed in Fig. 3.1.

This study assumes that both large scale path loss and shadow fading are compensated by

the perfect power control mechanism. Let I

ji,g

be the average inter-cell interference that all

users n

j,g

using services g with activity factor v

g

and received signal S

g

at BS j impose on

BS i. Modifying the average inter-cell interference given by [5], the equation becomes [43]

I

(g)

ji

= S

g

v

g

n

j,g

e

(γσs)

2

A

j

_ _

C

j

r

m

j

(x, y)

r

m

i

(x, y)

w(x, y) dA(x, y), (3.1)

where γ = ln(10)/10, σ

s

is the standard deviation of the attenuation for the shadow fading,

m is the path loss exponent, and w(x, y) is the user distribution density at (x, y). Let κ

ji,g

be the per-user (with service g) relative inter-cell interference factor from cell j to BS i,

κ

ji,g

=

e

(γσs)

2

A

j

_ _

C

j

r

m

j

(x, y)

r

m

i

(x, y)

w(x, y) dA(x, y). (3.2)

The inter-cell interference density I

inter

ji

from cell j to BS i from all services G becomes

I

inter

ji

=

1

W

G

g=1

I

(g)

ji

, (3.3)

19

where W is the bandwidth of the system. Eq. (3.3) can be rewritten as

I

inter

ji

=

1

W

G

g=1

S

g

v

g

n

j,g

κ

ji,g

. (3.4)

Thus, the total inter-cell interference density I

inter

i

from all other cells to BS i is

I

inter

i

=

1

W

M

j=1,j=i

G

g=1

S

g

v

g

n

j,g

κ

ji,g

, (3.5)

where M is the total number of cells in the network.

If the user distribution density can be approximated, then, κ

ji,g

needs to be calculated

only once. The user distribution is modeled with a 2-dimensional Gaussian function as

follows

w(x, y) =

η

2πσ

1

σ

2

e

−

1

2

(

x−µ

1

σ

1

)

2

e

−

1

2

(

y−µ

2

σ

2

)

2

, (3.6)

where η is a user density normalizing parameter.

By specifying the means µ

1

and µ

2

and the standard deviations σ

1

and σ

2

of the distri-

bution for every cell, an approximation can be found for a wide range of user distributions

ranging from uniform to hot-spot clusters. These results are compared with simulations to

determine the value of η experimentally.

3.4 Numerical Results

The results shown are for a twenty-seven cell network topology used in [5, 6]. The COST-231

propagation model with a carrier frequency of 1800 MHz, average base station height of 30

meters and average mobile height of 1.5 meters, is used to determine the coverage region.

The path loss coeﬃcient m is 4. The shadow fading standard deviation σ

s

is 6 dB. We

assume only one service, i.e., G = 1. The processing gain

W

R

is 21.1 dB. The activity factor,

v, is 0.375.

The simulator used for comparison is an extension of the software tools CDMA Capacity

Allocation and Planning (CCAP) [3]. CCAP, written in MATLAB, was developed at Wash-

ington University in St. Louis for numerical analysis of optimization techniques developed

20

in [5] to compute the capacity of CDMA networks. This study extends CCAP for WCDMA

networks and uses the 2-dimensional Gaussian function for w(x, y). The following models

show that by using 2-D Gaussian distribution, many diﬀerent scenarios can be modeled,

including users uniformly distributed, users clustered at the center of the cells, and users at

the cells’ boundaries. The results are veriﬁed with [6], where actual distances were used to

simulate real-time users entering the network for the calculation of interference.

3.4.1 Uniform Distribution of Users

The network with diﬀerent values of σ

1

and σ

2

has been analyzed, while keeping µ

1

and µ

2

equal to zero in (3.6). Table 3.1 shows the maximum number of users in every cell for the 27

cell WCDMA network, as the values of σ

1

and σ

2

are increased from 5000 to 15000, while µ

1

= 0 and µ

2

= 0. These increments of σ

1

, σ

2

result in users spread out, almost uniformly in

the cells. Fig. 3.2 shows the 2-D Gaussian approximation of users uniformly distributed in

the cells with σ

1

= σ

2

= 12000. The total number of users is 548. This compares well with

simulation results presented in Fig. 3.3, which yields a total number of users equal to 554

when they are placed uniformly in the cells.

3.4.2 Users Densely Clustered at the Center of the Cells

Table 3.2 shows the maximum number of users in every cell for the 27-cell WCDMA network

as the values of σ

1

and σ

2

are increased from 100 to 400 while µ

1

= 0 and µ

2

= 0. This results

in users densely clustered around the BSs. Fig. 3.4 shows the 2-D Gaussian approximation

with σ

1

= σ

2

= 100. The maximum number of users is 1026. This compares exactly with

simulation results presented in Fig. 3.5, which also yields a total number of users equal to

1026. In this conﬁguration, the users cause the least amount of interference to the network

by reducing the power gain required to maintain a desired signal-to-noise ratio.

21

Table 3.1: The maximum number of users in every cell for the 27 cell WCDMA network

(with σ

1

and σ

2

are increased from 5000 to 15000 while µ

1

= 0 and µ

2

= 0). This results in

users distributed uniformly in all BSs.

σ = σ

1

, σ

2

5000 7000 10000 12000 15000 Capacity from [5]

Cell

1

18 18 18 18 18 18

Cell

2

18 18 18 18 18 18

Cell

3

18 18 18 17 17 17

Cell

4

18 18 18 17 17 17

Cell

5

18 18 18 18 18 18

Cell

6

18 18 18 17 17 17

Cell

7

18 18 18 17 17 17

Cell

8

18 18 18 18 18 18

Cell

9

18 17 17 17 17 17

Cell

10

22 21 21 21 21 21

Cell

11

22 22 22 21 21 21

Cell

12

22 21 21 21 21 21

Cell

13

17 17 17 17 17 17

Cell

14

18 18 18 18 18 18

Cell

15

18 17 17 17 17 17

Cell

16

22 21 21 21 21 21

Cell

17

22 22 21 21 21 21

Cell

18

22 21 21 21 21 21

Cell

19

18 17 17 17 17 17

Cell

20

25 25 25 25 25 25

Cell

21

25 25 24 24 24 24

Cell

22

25 25 24 24 24 24

Cell

23

25 25 25 25 25 25

Cell

24

25 25 25 25 25 25

Cell

25

25 25 24 24 24 24

Cell

26

25 25 24 24 24 24

Cell

27

25 25 25 25 25 25

Total Users 565 558 553 548 548 548

22

Figure 3.2: 2-D Gaussian approximation of users uniformly distributed in the cells. σ

1

= σ

2

= 12000, µ

1

= µ

2

= 0. The maximum number of users is 548.

Figure 3.3: Simulated network capacity where users are uniformly distributed in the cells.

The maximum number of users is 554.

23

Table 3.2: The maximum number of users in 27 cells of WCDMA network as the values of

σ

1

and σ

2

are increased from 100 to 400 while µ

1

= 0 and µ

2

= 0. This results in users

densely clustered around the BSs.

σ = σ

1

, σ

2

σ = 100 σ = 200 σ = 300 σ = 400

Cell

1

38 38 37 34

Cell

2

38 38 37 34

Cell

3

38 38 37 35

Cell

4

38 38 37 35

Cell

5

38 38 37 34

Cell

6

38 38 37 35

Cell

7

38 38 37 35

Cell

8

38 38 37 35

Cell

9

38 38 37 35

Cell

10

38 38 37 36

Cell

11

38 38 37 36

Cell

12

38 38 37 36

Cell

13

38 38 37 35

Cell

14

38 38 37 35

Cell

15

38 38 37 35

Cell

16

38 38 37 35

Cell

17

38 38 37 35

Cell

18

38 38 37 35

Cell

19

38 38 37 35

Cell

20

38 38 37 36

Cell

21

38 38 38 36

Cell

22

38 38 38 37

Cell

23

38 38 38 36

Cell

24

38 38 38 36

Cell

25

38 38 37 36

Cell

26

38 38 37 36

Cell

27

38 38 37 36

Total Users 1026 1026 1003 954

24

Figure 3.4: 2-D Gaussian approximation of users densely clustered around the BSs. σ

1

= σ

2

= 100, µ

1

= µ

2

= 0. The maximum number of users is 1026.

Figure 3.5: Simulated network capacity where users are densely clustered around the BSs

causing the least amount of inter-cell interference. The maximum number of users is 1026

in the network.

25

Figure 3.6: 2-D Gaussian approximation of users clustered at the boundaries of the cells.

The values of σ

1

, σ

2

, µ

1

, and µ

2

may be diﬀerent in the diﬀerent cells and are given in Table

3.3. The maximum number of users is 133.

3.4.3 Users Distributed at Cells’ Boundaries

Fig. 3.6 shows the 2-D Gaussian approximation of users clustered at the boundaries of the

cells. The values of σ

1

, σ

2

, µ

1

, and µ

2

may be diﬀerent in the diﬀerent cells and are given

in Table 3.3. The maximum number of users is 133. These results are close to what was

attained through simulation. The maximum network capacity was purposely decreased by

the simulator by placing the users so they caused the maximum interference to the network.

The simulation yielded a total capacity of 108 users, with only 4 users in each cell. The

pattern seen in Fig. 3.7 shows that the simulator placed the users at the extreme corners

of their respective cells. The placement at extremities would require users to increase their

power gain causing much more interference to other users.

26

Table 3.3: The values of σ

1

, σ

2

, µ

1

, and µ

2

for the 2-D Gaussian approximation of users

clustered at the boundaries of the cells as shown in Fig. 3.6. The maximum number of users

is 133.

µ

1

σ

1

µ

2

σ

2

Cell

1

-1400 300 -900 300

Cell

2

-1400 300 800 300

Cell

3

-1400 300 800 300

Cell

4

0 300 -1700 300

Cell

5

0 300 -1600 300

Cell

6

1300 300 -800 300

Cell

7

-1400 300 900 300

Cell

8

-1300 300 900 300

Cell

9

0 300 1500 300

Cell

10

0 300 1600 300

Cell

11

0 300 1550 300

Cell

12

-1400 300 900 300

Cell

13

0 300 1500 300

Cell

14

1300 300 900 300

Cell

15

1300 300 -800 300

Cell

16

-1350 300 -850 300

Cell

17

-1400 300 -900 300

Cell

18

0 300 -1600 300

Cell

19

-1400 300 -800 300

Cell

20

-1400 300 -800 300

Cell

21

-1350 300 800 300

Cell

22

0 300 1600 300

Cell

23

1350 300 800 300

Cell

24

1400 300 -800 300

Cell

25

0 300 -1700 300

Cell

26

0 300 -1600 300

Cell

27

-1350 300 -850 300

27

Figure 3.7: Simulated network capacity where users are clustered at the boundaries of the

cells causing the most amount of inter-cell interference. The maximum number of users is

only 108 in the network.

3.5 Conclusions

An analytical model has been presented for approximating the user distributions in multi-cell

WCDMA networks using 2-dimensional Gaussian function by determining the means and

the standard deviations of the distributions for every cell. This allowed for the calculation of

the inter-cell interference and the reverse-link capacity of the network. The model compares

well with simulation results and is fast and accurate enough to be used eﬃciently in the

planning process of large WCDMA networks.

28

CHAPTER 4

WCDMA CAPACITY

4.1 Introduction

3G cellular systems are identiﬁed as International Mobile Telecommunications-2000 under In-

ternational Telecommunication Union and as Universal Mobile Telecommunications Systems

(UMTS) by European Telecommunications Standards Institute. Besides voice capability in

2G, the new 3G systems are required to have additional support on a variety of data-rate

services using multiple access techniques. CDMA is the fastest-growing digital wireless tech-

nology since its ﬁrst commercialization in 1994. The major markets for CDMA are North

America, Latin America, and Asia (particularly Japan and Korea). In total, CDMA has

been adopted by more than 100 operators across 76 countries around the globe [16]. Ac-

cording to [25, 50], CDMA technology can oﬀer about 7 to 10 times the capacity of analog

technologies and up to 6 times the capacity of digital technologies such as TDMA. With

its tremendous advantages such as voice quality, system reliability, and handset battery life

compared to TDMA and FDMA technologies, WCDMA, the next generation of CDMA, is

the best candidate for 3G cellular systems [29, 45].

4.2 Related Work

Since the ﬁrst comparisons of multiple access schemes for UMTS [45], which found that

WCDMA was well suited for supporting variable bit rate services, several research on

WCDMA capacity has been considered.

In [79], the authors present a method to calculate the WCDMA reverse link Erlang

capacity based on the Lost Call Held (LCH) model as described in [69]. This algorithm

29

calculates the occupancy distribution and capacity of UMTS/WCDMA systems based on

a system outage condition. In this research, the authors derive a closed form expression

of Erlang capacity for a single type of traﬃc loading and compare analytical results with

simulations results.

The same LCH model was also used in [78] to calculate the forward link capacity of

UMTS/WCDMA systems based on the systemoutage condition. In the forward link, because

many users share the BS transmission power, the capacity is calculated at the BS. The

transmission power from the BS is provided to each user based on each user’s relative need.

The access in the calculation of forward link capacity is one-to-many rather than many-to-

one as in the reverse link. In this research, the authors provide capacity calculation results

and performance evaluation through simulation.

An alternate approach, where MSs are synchronized on the uplink, i.e., signals trans-

mitted from diﬀerent MSs are time aligned at the BS, has been considered. Synchronous

WCDMA looks at time synchronization for signal transmission between the BS and MS to

improve network capacity. The performance of an uplink-synchronous WCDMA is analyzed

in [14]. Scrambling codes are unique for each cell. MSs in the same cell share the same

scrambling code, while diﬀerent orthogonal channelization codes are derived from the set

of Walsh codes. In [14], the potential capacity gain is about 35.8% in a multicell scenario

with inﬁnite number of channelization codes per cell and no soft handoﬀ capability between

MSs and BSs. However, the capacity gain in a more realistic scenario is reduced to 9.6%

where soft handoﬀ is enabled. The goal of this uplink-synchronous method in WCDMA is to

reduce intra-cell interference. But the implementation is fairly complex while the potential

capacity gain is not very high.

In this work, we will calculate the maximum reverse link capacity in UMTS/WCDMA

systems for a given quality of service requirements.

30

4.3 WCDMA Capacity with Perfect Power Control

In WCDMA, with perfect power control (PPC) between BSs and MSs, the energy per bit to

total interference density at BS i for a service g is given by [43, 55]

_

E

b

I

0

_

i,g

=

Sg

Rg

N

0

+ I

inter

i

+ I

own

i

− S

g

v

g

, (4.1)

where N

0

is the thermal noise density, and R

g

is the bit rate for service g. I

inter

i

was calculated

in section 3.3. I

own

i

is the total intra-cell interference density caused by all users in cell i.

Thus I

own

i

is given by

I

own

i

=

1

W

G

g=1

S

g

v

g

n

i,g

. (4.2)

Let τ

g

be the minimum signal-to-noise ratio, which must received at a BS to decode the

signal of a user with service g, and S

∗

g

be the maximum signal power, which the user can

transmit. Substituting (3.5) and (4.2) into (4.1), we have for every cell i in the WCDMA

network, the number of users n

i,g

in BS i for a given service g needs to meet the following

inequality constraint

τ

g

≤

S

∗

g

Rg

N

0

+

S

∗

g

W

_

G

g=1

n

i,g

v

g

+

M

j=1,j=i

G

g=1

n

j,g

v

g

κ

ji,g

−v

g

_

, for i=1, ...,M. (4.3)

After rearranging terms, (4.3) can be rewritten as

G

g=1

n

i,g

v

g

+

M

j=1,j=i

G

g=1

n

j,g

v

g

κ

ji,g

−v

g

≤ c

(g)

eff

, for i=1, ...,M, (4.4)

where

c

(g)

eff

=

W

R

g

_

1

τ

g

−

R

g

S

∗

g

/N

0

_

. (4.5)

The capacity in a WCDMA network is deﬁned as the maximum number of simultaneous

users (n

1,g

, n

2,g

, ..., n

M,g

) for all services g = 1, ..., G that satisfy (4.4).

31

4.4 WCDMA Capacity with Imperfect Power Control

The calculation of WCDMA network capacity, which was formulated in section 4.3, assumes

perfect power control between the BSs and MSs. However, transmitted signals between BSs

and MSs are subject to multipath propagation conditions, which make the received

_

E

b

Io

_

i,g

signals vary according to a log-normal distribution with a standard deviation on the order

of 1.5 to 2.5 dB [69]. Thus, in the imperfect power control (IPC) case, the constant value

of (E

b

)

i,g

in each cell i for every user with service g needs to be replaced by the variable

(E

b

)

i,g

=

i,g

(E

b

)

o,g

, which is log-normally distributed. We deﬁne

x

i,g

= 10log

10

_

i,g

(E

b

)

o,g

I

0

_

, (4.6)

to be a normally distributed random variable with mean m

c

and standard deviation σ

c

.

Hence, inverting (4.6), we have

(E

b

)

o,g

I

0

i,g

= 10

x

i,g

/10

= e

βx

i,g

, (4.7)

where β = ln(10)/10.

According to [69], by evaluating the nth moment of

i,g

using the fact that x

i,g

is Gaussian

with mean m

c

and standard deviation σ

c

, then taking the average value (the expected value),

we have

E

_

10

(x

i,g

/10)

_

= E

_

(E

b

)

o,g

I

0

i,g

_

=

(E

b

)

o,g

I

0

E[

i,g

] = e

βmc

e

(βσc)

2

2

. (4.8)

We can choose

(E

b

)o,g

I

0

such that

(E

b

)

o,g

I

0

= e

βmc

= 10

mc/10

, (4.9)

which makes

(E

b

)

o,g

I

0

= median

_

_

E

b

I

0

_

i,g

_

, (4.10)

for every user with service g in cell i.

Thus, the expected value becomes

E

_

(E

b

)

o,g

I

0

i,g

_

=

(E

b

)

i,g

I

0

e

(βσc)

2

2

. (4.11)

32

Figure 4.1: Generation of OVSF codes for diﬀerent Spreading Factors.

As a result of (4.11), c

(g)

eff IPC

becomes c

(g)

eff

/ e

(βσc)

2

2

.

4.5 Spreading and Scrambling

Communication from a single source is separated by channelization codes, i.e., the dedicated

physical channel in the uplink and the downlink connections within one sector from one MS.

The Orthogonal Variable Spreading Factor (OVSF) codes, which were originally introduced

in [21], were used to be channelization codes for UMTS.

The use of OVSF codes allows the orthogonality and spreading factor (SF) to be changed

between diﬀerent spreading codes of diﬀerent lengths. Fig. 4.1 depicts the generation of

diﬀerent OVSF codes for diﬀerent SF values.

The data signal after spreading is then scrambled with a scrambling codes to separate

MSs and BSs from each other. Scrambling is used on top of spreading, thus it only makes

the signals from diﬀerent sources distinguishable from each other. Fig. 4.2 depicts the

relationship between the spreading and scrambling process. Table 4.1 describes the diﬀerent

functionality of the channelization and the scrambling codes.

33

Figure 4.2: Relationship between spreading and scrambling.

Table 4.1: Functionality of the channelization and scrambling codes.

Channelization code Scrambling code

Usage Uplink: Separation of physical data (DPDCH)

and control channels (DPCCH) from same MS

Downlink: Separation of downlink connections to

diﬀerent MSs within one cell.

Uplink: Separation of MSs

Downlink: Separation of sectors (cells)

Length Uplink: 4-256 chips same as SF

Downlink 4-512 chips same as SF

Uplink: 10 ms = 38400 chips

Downlink: 10 ms = 38400 chips

Number of codes Number of codes under one scrambling code =

spreading factor

Uplink: Several millions

Downlink: 512

Code family Orthogonal Variable Spreading Factor Long 10 ms code: Gold Code

Short code: Extended S(2) code family

Spreading Yes, increases transmission bandwidth No, does not aﬀect transmission bandwidth

34

Figure 4.3: 12.2 Kbps Uplink Reference channel.

The typical required data rate or Dedicated Traﬃc Channel (DTCH) for a voice user is

12.2 Kbps. However, the Dedicated Physical Data Channel (DPDCH), which is the actual

transmitted data rate, is dramatically increased due to the incorporated Dedicated Control

Channel (DCCH) information, and the processes of Channel Coding, Rate Matching, and

Radio Frame Alignment. Fig. 4.3 depicts the process of creating the actual transmitted

signal for a voice user. Fig. 4.4 shows the DPDCH data rate requirement for 64 Kbps data

user. Table 4.2 shows the approximation of the maximum user data rate with

1

2

rate coding

for diﬀerent values of DPDCH.

4.6 Numerical Results

The results shown are for a twenty-seven cell network topology used in [2, 5, 6, 43]. The

COST-231 propagation model with a carrier frequency of 1800 MHz, average base station

height of 30 meters and average mobile height of 1.5 meters, is used to determine the coverage

region. The path loss coeﬃcient m is 4. The shadow fading standard deviation σ

s

is 6 dB.

The processing gain

W

Rg

is 6.02 dB, 12.04 dB, 18.06 dB, and 24.08 dB for Spreading Factor

35

Figure 4.4: 64 Kbps Uplink Reference channel.

Table 4.2: Uplink DPDCH data rates.

DPDCH Spreading Factor DPDCH channel bit rate (Kbps) Maximum user data rate with

1

2

rate coding (approx.)

256 15 7.5 Kbps

128 30 15 Kbps

64 60 30 Kbps

32 120 60 Kbps

16 240 120 Kbps

8 480 240 Kbps

4 960 480 Kbps

4, with 6 parallel codes 5740 2.8 Mbps

36

equal to 4, 16, 64, and 256, respectively. The activity factor, v, is 0.375.

This study extends the software tools (CCAP) [3] to analyze diﬀerent SF in WCDMA

networks. CCAP, written in MATLAB, was developed at Washington University in St. Louis

for numerical analysis of optimization techniques developed in [5] to compute the capacity

of CDMA networks.

The WCDMA network with 27 omni-directional antenna cells (1 sector per cell) was

analyzed for evaluation of capacity using user modeling with the 2-D Gaussian function,

whose parameters were determined in Chap. 3, and traditional methods of modeling uniform

user distribution. The network with diﬀerent values for

E

b

I

0

was analyzed for diﬀerent SF

values of 4, 16, 64, and 256.

4.6.1 WCDMA Capacity Optimization with SF of 256

First, we set SF to 256, which is used to carry data for the control channel. Table 4.3 shows

the number of slots per cell with omni-directional antenna for the scenario with

E

b

I

0

= 7.5 dB

while the standard deviation of the imperfect power control is increased from 0 to 2.5 dB.

Fig. 4.5 shows the optimized average number of slots per sector for the 27 cells WCDMA

network as the

E

b

I

0

is increased from 5 dB to 10 dB and the standard deviation of imperfect

power control is increased from 0 to 2.5 dB. Because of IPC, to get the same average number

of slots per sector as PPC, we have to decrease the SIR threshold by 0.5 dB to 1.5 dB. Fig.

4.5 also shows that the traditional uniform user distribution modeling matches well with the

2-D Gaussian model.

4.6.2 WCDMA Capacity Optimization with SF of 64

Next, we set SF to 64, which is used for voice communication as shown in Fig. 4.3. As a

result of lowering the SF to 64, the number of slots per sector decreases by almost a factor

of 4 compared to SF equal 256 (from 60.58 to 15.56 slots when

E

b

Io

= 7.5 dB in PPC). Table

37

Table 4.3: Capacity calculation for uniform user distribution with SF = 256 and

E

b

Io

= 7.5

dB.

User Modeling with 2-D Gaussian function Uniform User Distribution

Imperfect Power Control. σ: 0.0 1.5 2.5 0.0 1.5 2.5

c

eff

110.20 97.92 79.40 110.20 97.92 79.40

Cell

1

52.86 46.97 38.09 52.67 46.80 37.95

Cell

2

53.95 47.94 38.87 53.79 47.80 38.75

Cell

3

51.84 46.07 37.35 51.64 45.88 37.20

Cell

4

51.84 46.07 37.35 51.64 45.88 37.20

Cell

5

53.95 47.94 38.87 53.79 47.80 38.75

Cell

6

51.85 46.07 37.36 51.64 45.88 37.20

Cell

7

51.84 46.07 37.35 51.64 45.88 37.20

Cell

8

53.00 47.10 38.19 52.74 46.86 38.00

Cell

9

50.73 45.08 36.55 50.49 44.87 36.38

Cell

10

62.74 55.75 45.20 62.51 55.54 45.04

Cell

11

63.29 56.24 45.60 63.08 56.05 45.45

Cell

12

62.73 55.74 45.20 62.51 55.54 45.04

Cell

13

50.73 45.08 36.55 50.49 44.87 36.38

Cell

14

53.01 47.10 38.19 52.74 46.86 38.00

Cell

15

50.73 45.08 36.55 50.49 44.87 36.38

Cell

16

62.71 55.72 45.18 62.51 55.54 45.04

Cell

17

63.27 56.22 45.59 63.08 56.05 45.45

Cell

18

62.71 55.73 45.19 62.51 55.54 45.04

Cell

19

50.74 45.08 36.56 50.49 44.87 36.38

Cell

20

73.40 65.22 52.88 73.26 65.10 52.78

Cell

21

71.84 63.84 51.76 71.65 63.66 51.62

Cell

22

71.86 63.85 51.77 71.65 63.66 51.62

Cell

23

73.43 65.24 52.90 73.26 65.10 52.78

Cell

24

73.43 65.25 52.91 73.26 65.10 52.78

Cell

25

71.83 63.83 51.75 71.65 63.66 51.62

Cell

26

71.82 63.81 51.74 71.65 63.66 51.62

Cell

27

73.40 65.22 52.89 73.26 65.10 52.78

Network Capacity 1635.54 1453.29 1178.41 1630.06 1448.42 1174.46

Average Capacity per Sector 60.58 54.09 43.64 60.37 53.65 43.50

38

20 40 60 80 100 120

5

6

7

8

9

10

Average number of slots per sector

S

I

R

t

h

r

e

s

h

h

o

l

d

i

n

(

d

B

)

SF = 256

PPC (Gaussian), σ = 0 dB

IPC (Gaussian), σ = 1.5 dB

IPC (Gaussian), σ = 2.5 dB

PPC (Uniform), σ = 0 dB

IPC (Uniform), σ = 1.5 dB

IPC (Uniform), σ = 2.5 dB

Figure 4.5: Average number of slot per sector for perfect and imperfect power control analysis

with a Spreading Factor of 256.

4.4 shows the number of slots per cell for the scenario with

E

b

I

0

= 7.5 dB while the standard

deviation of the imperfect power control is increased from 0 to 2.5 dB. Fig. 4.6 shows the

optimized average number of slots per sector for the 27 cells WCDMA network as the

E

b

I

0

is increased from 5 dB to 10 dB and the standard deviation of imperfect power control is

increased from 0 to 2.5 dB. Because of IPC, to get the same amount of the average number

of slots per sector as PPC, the SIR threshold has to be decreased by 0.5 dB to 1.5 dB. Fig.

4.6 also shows that the traditional uniform user distribution modeling matches well with the

2-D Gaussian model.

4.6.3 WCDMA Capacity Optimization with SF of 16

Next, we set SF to 16, which is used for 64 Kbps data communication as shown in Fig. 4.4.

As a result of lowering the SF to 16, the number of slots per sector decreases by almost a

factor of 4 compared to SF equal 64 (from 15.56 to 4.30 slots when

E

b

Io

= 7.5 dB in PPC).

39

Table 4.4: Capacity calculation for uniform user distribution with SF = 64 and

E

b

Io

= 7.5 dB.

User Modeling with 2-D Gaussian function Uniform User Distribution

Imperfect Power Control. σ: 0.0 1.5 2.5 0.0 1.5 2.5

c

eff

28.30 25.23 20.60 28.30 25.23 20.60

Cell

1

13.58 12.10 9.88 13.53 12.06 9.85

Cell

2

13.86 12.35 10.09 13.82 12.32 10.06

Cell

3

13.32 11.87 9.69 13.26 11.82 9.65

Cell

4

13.32 11.87 9.69 13.26 11.82 9.65

Cell

5

13.86 12.35 10.09 13.82 12.32 10.06

Cell

6

13.32 11.87 9.69 13.26 11.82 9.65

Cell

7

13.32 11.87 9.69 13.26 11.82 9.65

Cell

8

13.61 12.14 9.91 13.54 12.08 9.86

Cell

9

13.03 11.62 9.48 12.97 11.56 9.44

Cell

10

16.11 14.37 11.73 16.06 14.31 11.69

Cell

11

16.26 14.49 11.83 16.20 14.44 11.79

Cell

12

16.11 14.36 11.73 16.06 14.31 11.69

Cell

13

13.03 11.62 9.48 12.97 11.56 9.44

Cell

14

13.62 12.14 9.91 13.54 12.08 9.86

Cell

15

13.03 11.62 9.48 12.97 11.56 9.44

Cell

16

16.11 14.36 11.72 16.06 14.31 11.69

Cell

17

16.25 14.49 11.83 16.20 14.44 11.79

Cell

18

16.11 14.36 11.72 16.06 14.31 11.69

Cell

19

13.03 11.62 9.49 12.97 11.56 9.44

Cell

20

18.85 16.81 13.72 18.82 16.78 13.70

Cell

21

18.45 16.45 13.43 18.40 16.41 13.39

Cell

22

18.46 16.45 13.43 18.40 16.41 13.39

Cell

23

18.86 16.81 13.73 18.82 16.78 13.70

Cell

24

18.86 16.81 13.73 18.82 16.78 13.70

Cell

25

18.45 16.45 13.43 18.40 16.41 13.39

Cell

26

18.45 16.44 13.43 18.40 16.41 13.39

Cell

27

18.85 16.81 13.72 18.82 16.78 13.70

Network Capacity 420.07 374.50 305.77 418.67 373.25 304.75

Average Capacity per Sector 15.56 13.87 11.32 15.51 13.82 11.29

40

5 10 15 20 25 30

5

6

7

8

9

10

Average number of slots per sector

S

I

R

t

h

r

e

s

h

h

o

l

d

i

n

(

d

B

)

SF = 64

PPC (Gaussian), σ = 0 dB

IPC (Gaussian), σ = 1.5 dB

IPC (Gaussian), σ = 2.5 dB

PPC (Uniform), σ = 0 dB

IPC (Uniform), σ = 1.5 dB

IPC (Uniform), σ = 2.5 dB

Figure 4.6: Average number of slot per sector for perfect and imperfect power control analysis

with a Spreading Factor of 64.

Table 4.5 shows the number of slots per cell for the scenario with

E

b

I

0

= 7.5 dB while the

standard deviation of the imperfect power control is increased from 0 to 2.5 dB. Fig. 4.7

shows the optimized average number of slots per sector for the 27 cells WCDMA network

as the

E

b

I

0

is increased from 5 dB to 10 dB and the standard deviation of imperfect power

control is increased from 0 to 2.5 dB. Because of IPC, to get the same average number of

slots per sector as PPC, we have to increase the SIR threshold by 0.5 dB to 1.5 dB. Fig.

4.7 also shows that the traditional uniform user distribution modeling matches well with the

2-D Gaussian model.

4.6.4 WCDMA Capacity Optimization with SF of 4

Next, we set SF to 4, which is used for 256 Kbps data communication between BSs and MSs.

As a result of lowering the SF to 4, the number of slots per sector decreases signiﬁcantly to

1.49 while keeping

E

b

Io

= 7.5 dB in PPC. Table 4.6 shows the number of slots per cell for the

41

Table 4.5: Capacity calculation for uniform user distribution with SF = 16 and

E

b

Io

= 7.5 dB.

User Modeling with 2-D Gaussian function Uniform User Distribution

Imperfect Power Control. σ: 0.0 1.5 2.5 0.0 1.5 2.5

c

eff

7.83 7.06 5.90 7.83 7.06 5.90

Cell

1

3.75 3.39 2.83 3.74 3.37 2.82

Cell

2

3.83 3.46 2.89 3.82 3.45 2.88

Cell

3

3.68 3.32 2.78 3.67 3.31 2.77

Cell

4

3.68 3.32 2.78 3.67 3.31 2.77

Cell

5

3.83 3.46 2.89 3.82 3.45 2.88

Cell

6

3.68 3.32 2.78 3.67 3.31 2.77

Cell

7

3.68 3.32 2.78 3.67 3.31 2.77

Cell

8

3.76 3.40 2.84 3.75 3.38 2.82

Cell

9

3.60 3.25 2.72 3.59 3.23 2.70

Cell

10

4.46 4.02 3.36 4.44 4.00 3.35

Cell

11

4.50 4.05 3.39 4.48 4.04 3.38

Cell

12

4.46 4.02 3.36 4.44 4.00 3.35

Cell

13

3.60 3.25 2.72 3.59 3.23 2.70

Cell

14

3.77 3.40 2.84 3.75 3.38 2.82

Cell

15

3.60 3.25 2.72 3.59 3.23 2.70

Cell

16

4.45 4.02 3.36 4.44 4.00 3.35

Cell

17

4.49 4.05 3.39 4.48 4.04 3.38

Cell

18

4.45 4.02 3.36 4.44 4.00 3.35

Cell

19

3.60 3.25 2.72 3.59 3.23 2.70

Cell

20

5.21 4.70 3.93 5.20 4.69 3.92

Cell

21

5.10 4.60 3.85 5.09 4.59 3.84

Cell

22

5.10 4.60 3.85 5.09 4.59 3.84

Cell

23

5.22 4.70 3.93 5.20 4.69 3.92

Cell

24

5.22 4.70 3.93 5.20 4.69 3.92

Cell

25

5.10 4.60 3.85 5.09 4.59 3.84

Cell

26

5.10 4.60 3.85 5.09 4.59 3.84

Cell

27

5.21 4.70 3.93 5.20 4.69 3.92

Network Capacity 116.16 104.77 87.59 115.77 104.42 87.29

Average Capacity per Sector 4.30 3.88 3.24 4.29 3.87 3.23

42

0 2 4 6 8 10

5

6

7

8

9

10

Average number of slots per sector

S

I

R

t

h

r

e

s

h

h

o

l

d

i

n

(

d

B

)

SF = 16

PPC (Gaussian), σ = 0 dB

IPC (Gaussian), σ = 1.5 dB

IPC (Gaussian), σ = 2.5 dB

PPC (Uniform), σ = 0 dB

IPC (Uniform), σ = 1.5 dB

IPC (Uniform), σ = 2.5 dB

Figure 4.7: Average number of slot per sector for perfect and imperfect power control analysis

with a Spreading Factor of 16.

scenario with

E

b

I

0

= 7.5 dB while the standard deviation of the imperfect power control is

increased from 0 to 2.5 dB. Fig. 4.8 shows the optimized average number of slots per sector

for the 27 cells WCDMA network as the

E

b

I

0

is increased from 5 dB to 10 dB and the standard

deviation of imperfect power control is increased from 0 to 2.5 dB. The results matches the

previous trends for higher spreading factors.

4.7 Conclusions

An analytical model has been presented for optimizing capacity in multi-cell WCDMA net-

works. Numerical results show that the SIR threshold for the received signals is decreased

by 0.5 to 1.5 dB due to the imperfect power control. As expected, we can have many low

rate voice users or fewer data users as the data rate increases. The results also show that the

determined parameters of the 2-dimensional Gaussian model matches well with traditional

43

Table 4.6: Capacity calculation for uniform user distribution with SF = 4 and

E

b

Io

= 7.5 dB.

User Modeling with 2-D Gaussian function Uniform User Distribution

Imperfect Power Control. σ: 0.0 1.5 2.5 0.0 1.5 2.5

c

eff

2.71 2.52 2.23 2.71 2.52 2.23

Cell

1

1.30 1.21 1.07 1.29 1.20 1.06

Cell

2

1.33 1.23 1.09 1.32 1.23 1.09

Cell

3

1.27 1.18 1.05 1.27 1.18 1.04

Cell

4

1.27 1.18 1.05 1.27 1.18 1.04

Cell

5

1.33 1.23 1.09 1.32 1.23 1.09

Cell

6

1.27 1.18 1.05 1.27 1.18 1.04

Cell

7

1.27 1.18 1.05 1.27 1.18 1.04

Cell

8

1.30 1.21 1.07 1.30 1.20 1.07

Cell

9

1.25 1.16 1.02 1.24 1.15 1.02

Cell

10

1.54 1.43 1.27 1.54 1.43 1.26

Cell

11

1.55 1.44 1.28 1.55 1.44 1.27

Cell

12

1.54 1.43 1.27 1.54 1.43 1.26

Cell

13

1.25 1.16 1.02 1.24 1.15 1.02

Cell

14

1.30 1.21 1.07 1.30 1.20 1.07

Cell

15

1.25 1.16 1.02 1.24 1.15 1.02

Cell

16

1.54 1.43 1.27 1.54 1.43 1.26

Cell

17

1.55 1.44 1.28 1.55 1.44 1.27

Cell

18

1.54 1.43 1.27 1.54 1.43 1.26

Cell

19

1.25 1.16 1.02 1.24 1.15 1.02

Cell

20

1.80 1.68 1.48 1.80 1.67 1.48

Cell

21

1.76 1.64 1.45 1.76 1.64 1.45

Cell

22

1.77 1.64 1.45 1.76 1.64 1.45

Cell

23

1.80 1.68 1.48 1.80 1.67 1.48

Cell

24

1.80 1.68 1.48 1.80 1.67 1.48

Cell

25

1.76 1.64 1.45 1.76 1.64 1.45

Cell

26

1.76 1.64 1.45 1.76 1.64 1.45

Cell

27

1.80 1.68 1.48 1.80 1.67 1.48

Network Capacity 40.18 37.33 33.03 40.04 37.20 32.92

Average Capacity per Sector 1.49 1.38 1.22 1.48 1.38 1.22

44

0.5 1 1.5 2 2.5 3 3.5 4 4.5

5

6

7

8

9

10

Average number of slots per sector

S

I

R

t

h

r

e

s

h

h

o

l

d

i

n

(

d

B

)

SF = 4

PPC (Gaussian), σ = 0 dB

IPC (Gaussian), σ = 1.5 dB

IPC (Gaussian), σ = 2.5 dB

PPC (Uniform), σ = 0 dB

IPC (Uniform), σ = 1.5 dB

IPC (Uniform), σ = 2.5 dB

Figure 4.8: Average number of slot per sector for perfect and imperfect power control analysis

with a Spreading Factor of 4.

methods for modeling uniform user distribution. Our method of optimizing capacity is fast,

accurate, and can be implemented for large multi-cell WCDMA networks.

45

CHAPTER 5

WCDMA CALL ADMISSION CONTROL AND THROUGHPUT

5.1 Introduction

Call admission control (CAC) algorithms, in general, decide whether a new connection could

be admitted without impacting the quality of service (QoS) of current connections in a

network. In addition, CAC algorithms must be designed to fulﬁll a grade of service (GoS),

i.e., call blocking rate. The existence of CAC algorithms in mobile phone systems protects

the cellular network and users while achieving network performance objectives.

A CAC algorithm is considered global if the decision to admit a call is based on the

current total number of calls in the entire network, and local if the algorithm considers only

a single cell for making that decision. A global CAC is most often a centralized scheme while

a local CAC is distributed in nature. Each approach has advantages and disadvantages. In

this chapter, we design a CAC algorithm for multi-cell WCDMA networks and analyze its

complexity, implementation, and performance.

5.2 Feasible States

Recall from section 4.3 that the number of calls in every cell must satisfy

G

g=1

n

i,g

v

g

+

M

j=1,j=i

G

g=1

n

j,g

v

g

κ

ji,g

−v

g

≤ c

(g)

eff

for i=1..M, (5.1)

where

c

(g)

eff

=

W

R

g

_

1

τ

g

−

R

g

S

∗

g

/N

0

_

. (5.2)

A set of calls n =

_

¸

_

n

1,1

... n

1,G

... ... ...

n

M,1

... n

M,G

_

¸

_ satisfying the above equations is said to be in feasible

call conﬁguration or a feasible state, which meet the

E

b

I

0

constraint.

46

Denote by Ω the set of feasible states. Deﬁne the set of blocking states for service g in

cell i as

B

i,g

=

_

¸

¸

¸

¸

¸

¸

_

¸

¸

¸

¸

¸

¸

_

n ∈ Ω :

_

¸

¸

¸

¸

¸

¸

_

n

1,1

... n

1,g

... n

1,G

... ... ... ... ...

n

i,1

... n

i,g

... n

i,G

... ... ... ... ...

n

M,1

... n

M,g

... n

M,G

_

¸

¸

¸

¸

¸

¸

_

∈ Ω

_

¸

¸

¸

¸

¸

¸

_

¸

¸

¸

¸

¸

¸

_

. (5.3)

If a new connection or a handoﬀ connection with the service g arrives to cell i, it is blocked

when the current state of the network, n, is in B

i,g

.

5.3 Mobility Model

The call arrival process with service g to cell i is assumed to be a Poisson process with rate

λ

i,g

independent of other call arrival processes. The call dwell time is a random variable with

exponential distribution having mean 1/µ, and it is independent of earlier arrival times, call

durations and elapsed times of other users. At the end of a dwell time a call may stay in

the same cell, attempt a handoﬀ to an adjacent cell, or leave the network. Deﬁne q

ii,g

as the

probability that a call with service g in progress in cell i remains in cell i after completing

its dwell time. In this case, a new dwell time that is independent of the previous dwell time

begins immediately. Let q

ij,g

be the probability that a call with service g in progress in cell i

after completing its dwell time goes to cell j. If cells i and j are not adjacent, then q

ij,g

= 0

(∀ g ∈ G). We denote by q

i,g

the probability that a call with service g in progress in cell i

departs from the network.

This mobility model is attractive because we can easily deﬁne diﬀerent mobility scenarios

by varying the values of these probability parameters [65]. For example, if q

i,g

is constant

for all i and g, then the average dwell time of a call of the same service in the network

will be constant regardless of where the call originates and what the values of q

ii,g

and q

ij,g

are. Thus, by varying q

ii,g

’s and q

ij,g

’s for a service g, we can obtain low and high mobility

scenarios and compare the eﬀect of mobility on network attributes (e.g., throughput).

47

We assume that the occupancy of the cells evolves according to a birth-death process,

where the total arrival rate or oﬀered traﬃc for service g to cell i is ρ

i,g

, and the departure

rate from cell i when the network is in state n is n

i,g

µ

i,g

= n

i,g

µ(1 − q

ii,g

). Let ρ be the

matrix of oﬀered traﬃc of service g to the cells, µ the matrix of departure rates, and let

p(ρ, µ, n) be the stationary probability that the network is in state n. The distribution p is

obtained as follows

p(ρ, µ, n) =

_

¸

¸

¸

_

¸

¸

¸

_

P

0

M

k=1

G

g=1

(ρ

k,g

/µ

k,g

)

n

k,g

n

k,g

!

G

g=1

n

i,g

v

g

+

M

j=1,j=i

G

g=1

n

j,g

v

g

κ

ji,g

−v

g

≤ c

(g)

eff

,

for i = 1, ..., M,

0 otherwise,

(5.4)

where P

0

is a normalizing constant such that

n∈Ω

p(ρ, µ, n) = 1. The new call blocking

probability for service g in cell i, B

i,g

, is given by

B

i,g

=

n∈B

i,g

p(ρ, µ, n). (5.5)

This is also the blocking probability of handoﬀ calls due to the fact that handoﬀ calls and

new calls are treated in the same way by the network.

Let A

i

be the set of cells adjacent to cell i. Let ν

ji,g

be the handoﬀ rate out of cell j

oﬀered to cell i for service g. ν

ji,g

is the sum of the proportion of new calls of service g

accepted in cell j that go to cell i and the proportion of handoﬀ calls with service g accepted

from cells adjacent to cell j that go to cell i. Thus

ν

ji,g

= λ

j,g

(1 −B

j,g

)q

ji,g

+ (1 −B

j,g

)q

ji,g

x∈A

j

ν

xj,g

. (5.6)

Equation (5.6) can be rewritten as

ν

ji,g

= ν(B

j,g

, ρ

j,g

, q

ji,g

) = (1 −B

j,g

)q

ji,g

ρ

j,g

, (5.7)

where ρ

j,g

, the total oﬀered traﬃc to cell j for service g, is given by

ρ

j,g

= ρ(v, λ

j,g

, A

j

) = λ

j,g

+

x∈A

j

ν

xj,g

, (5.8)

48

and where v denotes the matrix whose components are the handoﬀ rates ν

ij

for i, j = 1, ...M.

The total oﬀered traﬃc can be obtained from a ﬁxed point model [32], which describes

the oﬀered traﬃc as a function of the handoﬀ rates and new call arrival rates, the handoﬀ

rates as a function of the blocking probabilities and the oﬀered traﬃc, and the blocking

probabilities as a function of the oﬀered traﬃc. For a given set of arrival rates, we use an

iterative method to solve the ﬁxed point equations. We deﬁne an initial value for the handoﬀ

rates. We calculate the oﬀered traﬃc by adding the given values of the arrival rates to the

handoﬀ rates. The blocking probabilities are now calculated using the oﬀered traﬃc. We then

calculate the new values of the handoﬀ rates and repeat. This approach has been extensively

utilized in the literature to obtain solutions of ﬁxed point problems [4, 37, 41, 66, 67, 68].

The questions of existence and uniqueness of the solution and whether the iterative approach

in fact converges to the solution (if a unique solution exists) are generally diﬃcult to answer

due to the complexity of the equations involved. Kelly has shown that for ﬁxed alternate

routing the solution to the ﬁxed point problem is in fact not unique [36]; in all the numerical

examples we solved, the iterative approach converged to a unique solution.

5.4 WCDMA Call Admission Control

A CAC algorithm can be constructed as follows. A call arriving to cell i with service g is

accepted if and only if the new state is a feasible state. Clearly this CAC algorithm requires

global state, i.e., the number of calls in progress in all the cells of the network. Furthermore,

to compute the blocking probabilities, the probability of each state in the feasible region

needs to be calculated. Since the cardinality of Ω is O(c

eff

M∗G

), the calculation of the

blocking probabilities has a computational complexity that is exponential in the number of

cells combined with number of available services.

In order to simplify the CAC algorithm, we consider only those CAC algorithms which

utilize local state, i.e., the number of calls in progress in the current cell. To this end we

49

deﬁne a state n to be admissible if

n

i,g

≤ N

i,g

for i = 1, ..., M and g = 1, ..., G, (5.9)

where N

i,g

is a parameter which denotes the maximum number of calls with service g allowed

to be admitted in cell i. Clearly the set of admissible states denoted Ω

**is a subset of the set
**

of feasible states Ω. The blocking probability for cell i with service g is then given by

B

i,g

= B(A

i,g

, N

i,g

) =

A

N

i,g

i,g

/N

i,g

!

N

i,g

k=0

A

k

i,g

/k!

, (5.10)

where A

i

= ρ

i,g

/µ

i,g

= ρ

i,g

/µ(1 −q

ii,g

) is the Erlang traﬃc in cell i with service g. We note

that the complexity to calculate the blocking probabilities in (5.10) is O(M ∗ G), and the

bit error rate requirement is guaranteed since Ω

⊂ Ω.

Once the maximum number of calls with diﬀerent service that are allowed to be admitted

in each cell, N, is calculated (this is done oﬄine and described in the next section), the CAC

algorithm for cell i for service g will simply compare the number of calls with service g

currently active in cell i to N

i,g

in order to accept or reject a new arriving call. Thus our

CAC algorithm is implemented with a computational complexity that is O(1).

5.5 Network Throughput

The throughput of cell i consists of two components: the new calls that are accepted in cell

i minus the forced termination due to handoﬀ failure of the handoﬀ calls into cell i for all

services g. Hence the total throughput, T, of the network is

T(B, ρ, λ) =

M

i=1

G

g=1

{λ

i,g

(1 −B

i,g

) −B

i,g

(ρ

i,g

−λ

i,g

)} , (5.11)

where B is the vector of blocking probabilities and λ is the matrix of call arrival rates.

To study the eﬀect of mobility and to diﬀerentiate between new calls and handoﬀ calls, the

throughput function can be generalized to a revenue function. The term revenue suggesting

50

an economic meaning is chosen to emphasize the rewards from not blocking a new call and

the penalty (whether measured monetarily or by customer aggravation) from having handoﬀ

calls blocked. Hence the revenue, H, becomes

H(B, ρ, λ) =

M

i=1

G

g=1

{r

i,g

λ

i,g

(1 −B

i,g

) −c

i,g

B

i,g

(ρ

i,g

−λ

i,g

)} , (5.12)

where r

i,g

is the revenue generated by accepting a new call with service g in cell i, and c

i,g

is

the cost of a forced termination of a call with service g due to a handoﬀ failure in cell i. The

values of r

i,g

and c

i,g

control the tradeoﬀ between new calls and handoﬀ calls. The choice of

r

i,g

and c

i,g

aﬀect the CAC algorithm through their eﬀect on N. For example assume that

r

i,g

= 1 and c

i,g

= c

g

for all i. Then, it is easy to see that for a given network topology

the choice of larger c

g

for service g will tend to increase the values of N for those cells into

which the handoﬀ rates are high while decreasing the values of N

i,g

for the other cells. These

are tools that allow the system administrator to place more importance on handoﬀ calls.

Surveys and market studies investigating these issues help network administrators set these

weights and achieve a desired tradeoﬀ.

5.6 Calculation of N

We formulate a constrained optimization problem in order to maximize the revenue subject

to upper bounds on the blocking probabilities and a lower bound on the signal-to-interference

constraints in (5.1). The goal is to optimize the utilization of network resources and provide

consistent GoS while at the same time maintaining the QoS, β

g

, for all the users for diﬀerent

services g. In this optimization problem the arrival rates are given and the maximum number

of calls that can be admitted in all the cells are the independent variables. This is given in

the following

51

max

N=

_

¸

_

N

1,1

... N

1,G

... ... ...

N

M,1

... N

M,G

_

¸

_

H(B, ρ, λ),

subject to B(A

i,g

, N

i,g

) ≤ β

g

,

G

g=1

N

i,g

v

g

+

M

j=1,j=i

G

g=1

N

j,g

v

g

κ

ji,g

−v

g

≤ c

(g)

eff

,

for i = 1, ..., M. (5.13)

The optimization problem in (5.13) is solved oﬄine to obtain the values of N.

5.7 Maximization of Throughput

A second optimization problem can be formulated in which the arrival rates and the maxi-

mum number of calls that can be admitted in all the cells are the independent variables and

the objective function is the throughput. This is given in the following

max

λ, N

T(B, ρ, λ),

subject to B(A

i,g

, N

i,g

) ≤ β

g

,

G

g=1

N

i,g

v

g

+

M

j=1,j=i

G

g=1

N

j,g

v

g

κ

ji,g

−v

g

≤ c

(g)

eff

,

for i = 1, ..., M. (5.14)

The optimized objective function of (5.14) provides an upper bound on the total throughput

that the network can carry. This is the network capacity for the given GoS and QoS.

5.8 Numerical Results

The following results have been obtained for a 27-cell CDMA network. The base stations are

located at the centers of a hexagonal cell whose radius is 1732 m. Base station 1 is located at

the center of the network; the base stations are numbered consecutively in a spiral pattern.

52

Table 5.1: The low mobility characteristics and parameters.

A

i

q

ij,g

q

ii,g

q

i,g

3 0.020 0.240 0.700

4 0.015 0.240 0.700

5 0.012 0.240 0.700

6 0.010 0.240 0.700

The COST-231 propagation model [49] with a carrier frequency of 1800 MHz, average base

station height of 30 m, and average mobile height of 1.5 m is used to determine the coverage

region. The path-loss coeﬃcient is 4. The shadow fading standard deviation is 6 dB and

the processing gain is 6.02 dB, 12.04 dB, 18.06 dB, and 24.08 dB for Spreading Factor of 4,

16, 64, and 256, respectively. The bit-energy-to-interference ratio threshold is 7.5 dB with

perfect power control for all four scenarios. The interference-to-background-noise ratio is 10

dB. The activity factor is 0.375.

Three mobility scenarios: no mobility, low mobility, and high mobility of users are con-

sidered. We assume that the mobility characteristics for a given service g stays the same

throughout diﬀerent cells in the network. The following parameters are used for the no

mobility case: q

ij,g

= 0, q

ii,g

= 0.3 and q

i,g

= 0.7 for all cells i and j. Tables 5.1 and 5.2

show respectively the mobility characteristics and parameters for the low and high mobility

cases. In all three mobility scenarios, the probability that a call leaves the network after

completing its dwell time is 0.7. Thus, regardless of where the call originates and mobility

scenario used, the average dwell time of a call in the network is constant. In the numerical

results below, for each SF value, we analyze the average throughput per cell by dividing

the results from (5.14) by the total number of cells in the network and multiplying by the

maximum data rate in Table 4.2.

53

Table 5.2: The high mobility characteristics and parameters.

A

i

q

ij,g

q

ii,g

q

i,g

3 0.1 0 0.700

4 0.075 0 0.700

5 0.060 0 0.700

6 0.050 0 0.700

• A

i

is the number of cells, which are adjacent to cell i.

• q

ii,g

is the probability that a call with service g in progress in cell i remains in cell i after completing its dwell time.

• q

ij,g

is the probability that a call with service g in progress in cell i after completing its dwell time goes to cell j.

• q

i,g

is the probability that a call with service g in progress in cell i departs from the network.

5.8.1 WCDMA Throughput Optimization with SF of 256

First, we set SF equal to 256, which is used to carry data for the control channel. Table

5.3 shows the optimized values of N for each cell for all three mobility models with perfect

power control and 2% blocking probability. Fig. 5.1 shows the optimized throughput per cell

for a blocking probability from 1% to 10%. The results for the average throughput for no

mobility and high mobility cases are almost identical while the throughput for low mobility

is higher for each blocking probability. The low mobility case has an equalizing eﬀect on

traﬃc resulting in slightly higher throughput.

5.8.2 WCDMA Throughput Optimization with SF of 64

Next, we set SF equal to 64, which is used for voice communication as shown in Fig. 4.3.

As a result of lowering the SF to 64, the number of possible concurrent connections within

one cell is also decreased. Because the throughput is calculated based on the number of

simultaneous connections between MSs and BSs, the lower trunking eﬃciency [48] leads to

54

Table 5.3: Calculation of N for uniform user distribution with SF = 256 and blocking

probability = 0.02.

No Mobility Low Mobility High Mobility

Cell ID N

i

N

i

N

i

Cell

1

52.86 52.86 52.86

Cell

2

53.95 53.95 53.95

Cell

3

51.84 51.84 51.84

Cell

4

51.84 51.84 51.84

Cell

5

53.95 53.95 53.95

Cell

6

51.85 51.85 51.85

Cell

7

51.85 51.85 51.85

Cell

8

53.00 53.00 53.00

Cell

9

50.73 50.73 50.73

Cell

10

62.74 62.74 62.74

Cell

11

63.29 63.29 63.29

Cell

12

62.73 62.73 62.73

Cell

13

50.73 50.73 50.73

Cell

14

53.01 53.01 53.01

Cell

15

50.73 50.73 50.73

Cell

16

62.71 62.71 62.71

Cell

17

63.27 63.27 63.27

Cell

18

62.71 62.71 62.71

Cell

19

50.74 50.74 50.74

Cell

20

73.40 73.40 73.40

Cell

21

71.84 71.84 71.84

Cell

22

71.86 71.86 71.86

Cell

23

73.43 73.43 73.43

Cell

24

73.43 73.43 73.43

Cell

25

71.83 71.83 71.83

Cell

26

71.82 71.82 71.82

Cell

27

73.40 73.40 73.40

55

0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1

250

255

260

265

270

275

280

285

290

Blocking probability threshold

A

v

e

r

a

g

e

T

h

r

o

u

g

h

p

u

t

p

e

r

C

e

l

l

(

k

b

p

s

p

e

r

u

n

i

t

t

i

m

e

)

SF=256

CAC high mobility

CAC low mobility

CAC no mobility

Figure 5.1: Average throughput in each cell for SF = 256.

lower throughput as shown in Fig. 5.2. Table 5.4 shows the optimized values of N for each

cell for all three mobility cases and SF equal to 64.

5.8.3 WCDMA Throughput Optimization with SF of 16

Next, we set SF equal to 16, which is used for 64 Kbps data communication as shown in

Fig. 4.4. As a result of lowering the SF to 16, the average number of slots within one cell

decreases to 4.30 as was shown in Table 4.5. The resulting throughput, as shown in Fig.

5.3, is much lower compared to the case with SF equal to 64 or 256. Table 5.5 shows the

optimized values of N for each cell for all three mobility cases with SF equal to 16.

5.8.4 WCDMA Throughput Optimization with SF of 4

Next, we set SF equal to 4, which is normally used for 256 Kbps data communication between

BSs and MSs. As a result of lowering the SF to 4, the average slots per sector decreases

signiﬁcantly to 1.49 with perfect power control and

E

b

I

0

= 7.5 dB as shown in Fig. 4.8. Table

56

0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1

180

190

200

210

220

230

240

250

Blocking probability threshold

A

v

e

r

a

g

e

T

h

r

o

u

g

h

p

u

t

p

e

r

C

e

l

l

(

k

b

p

s

p

e

r

u

n

i

t

t

i

m

e

)

)

SF=64

CAC high mobility

CAC low mobility

CAC no mobility

Figure 5.2: Average throughput in each cell for SF = 64.

0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1

90

100

110

120

130

140

150

160

170

Blocking probability threshold

A

v

e

r

a

g

e

T

h

r

o

u

g

h

p

u

t

p

e

r

C

e

l

l

(

k

b

p

s

p

e

r

u

n

i

t

t

i

m

e

)

SF=16

CAC high mobility

CAC low mobility

CAC no mobility

Figure 5.3: Average throughput in each cell for SF = 16.

57

Table 5.4: Calculation of N for uniform user distribution with SF = 64 and blocking prob-

ability = 0.02.

No Mobility Low Mobility High Mobility

Cell ID N

i

N

i

N

i

Cell

1

13.58 13.58 13.58

Cell

2

13.86 13.86 13.86

Cell

3

13.32 13.32 13.32

Cell

4

13.32 13.32 13.32

Cell

5

13.86 13.86 13.86

Cell

6

13.32 13.32 13.32

Cell

7

13.32 13.32 13.32

Cell

8

13.61 13.61 13.61

Cell

9

13.03 13.03 13.03

Cell

10

16.11 16.11 16.11

Cell

11

16.26 16.26 16.26

Cell

12

16.11 16.11 16.11

Cell

13

13.03 13.03 13.03

Cell

14

13.62 13.62 13.62

Cell

15

13.03 13.03 13.03

Cell

16

16.11 16.11 16.11

Cell

17

16.25 16.25 16.25

Cell

18

16.11 16.11 16.11

Cell

19

13.03 13.03 13.03

Cell

20

18.85 18.85 18.85

Cell

21

18.45 18.45 18.45

Cell

22

18.46 18.46 18.46

Cell

23

18.86 18.86 18.86

Cell

24

18.86 18.86 18.86

Cell

25

18.45 18.45 18.45

Cell

26

18.45 18.45 18.45

Cell

27

18.85 18.85 18.85

58

Table 5.5: Calculation of N for uniform user distribution with SF = 16 and blocking prob-

ability = 0.02.

No Mobility Low Mobility High Mobility

Cell ID N

i

N

i

N

i

Cell

1

3.75 3.75 3.75

Cell

2

3.83 3.83 3.83

Cell

3

3.68 3.68 3.68

Cell

4

3.68 3.68 3.68

Cell

5

3.83 3.83 3.83

Cell

6

3.68 3.68 3.68

Cell

7

3.68 3.68 3.68

Cell

8

3.76 3.76 3.76

Cell

9

3.60 3.60 3.60

Cell

10

4.46 4.46 4.46

Cell

11

4.50 4.50 4.50

Cell

12

4.46 4.46 4.46

Cell

13

3.60 3.60 3.60

Cell

14

3.77 3.77 3.77

Cell

15

3.60 3.60 3.60

Cell

16

4.45 4.45 4.45

Cell

17

4.49 4.49 4.49

Cell

18

4.45 4.45 4.45

Cell

19

3.60 3.60 3.60

Cell

20

5.21 5.21 5.21

Cell

21

5.10 5.10 5.10

Cell

22

5.10 5.10 5.10

Cell

23

5.22 5.22 5.22

Cell

24

5.22 5.22 5.22

Cell

25

5.10 5.10 5.10

Cell

26

5.10 5.10 5.10

Cell

27

5.21 5.21 5.21

59

0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1

20

30

40

50

60

70

80

90

Blocking probability threshold

A

v

e

r

a

g

e

T

h

r

o

u

g

h

p

u

t

p

e

r

C

e

l

l

(

k

b

p

s

p

e

r

u

n

i

t

t

i

m

e

)

SF=4

CAC high mobility

CAC low mobility

CAC no mobility

Figure 5.4: Average throughput in each cell for SF = 4.

5.5 shows the optimized values of N for each cell for all three mobility models with SF equal

to 4. The average throughput for all three mobility cases are almost identical as shown in

Fig. 5.4.

5.9 Conclusions

An analytical model has been presented for CAC algorithm for optimizing the throughput

in multi-cell WCDMA networks. For the high mobility case, every connection moves to a

new cell, which is equivalent to having new calls in every cell for each dwell time. Thus,

the throughput for the no mobility and high mobility scenarios are identical. However, the

throughput for the low mobility scenario is slightly higher because of the equalizing eﬀect

on traﬃc due to low mobility. Numerical results show that as the spreading factor increases,

the optimized throughput is better, due to the trunking eﬃciency for all the three mobility

models. Our CAC algorithm is computationally eﬃcient and can be implemented for large

multi-cell WCDMA networks.

60

Table 5.6: Calculation of N for uniform user distribution with SF = 4 and blocking proba-

bility = 0.02.

No Mobility Low Mobility High Mobility

Cell ID N

i

N

i

N

i

Cell

1

1.24 1.16 1.32

Cell

2

1.48 1.42 1.24

Cell

3

1.30 1.32 1.28

Cell

4

1.30 1.32 1.28

Cell

5

1.48 1.42 1.23

Cell

6

1.30 1.32 1.28

Cell

7

1.30 1.32 1.28

Cell

8

0.94 1.37 1.23

Cell

9

0.93 0.94 1.27

Cell

10

1.59 1.58 1.54

Cell

11

1.53 1.53 1.55

Cell

12

1.59 1.58 1.54

Cell

13

0.93 0.94 1.27

Cell

14

0.94 1.37 1.23

Cell

15

0.93 0.93 1.27

Cell

16

1.59 1.58 1.54

Cell

17

1.53 1.53 1.55

Cell

18

1.59 1.58 1.54

Cell

19

0.93 0.93 1.27

Cell

20

1.92 1.83 1.82

Cell

21

1.79 1.80 1.76

Cell

22

1.79 1.80 1.76

Cell

23

1.92 1.83 1.82

Cell

24

1.92 1.83 1.82

Cell

25

1.79 1.80 1.76

Cell

26

1.79 1.80 1.76

Cell

27

1.92 1.83 1.82

61

CHAPTER 6

CONCLUSIONS

6.1 Summary

Compared to other multiple access technologies like FDMA and TDMA, which divide the

total bandwidth to frequency sub-bands and time slots, respectively, CDMA is an unique

multiple access technology, where all users share the entire available bandwidth to transmit

and receive information. Interference generated by users in the network is the main limi-

tation for determining the network capacity. A large number of researchers in CDMA use

average interference for designing call admission control and calculating capacity. The ad-

vantage gained from this model over actual interference is a huge reduction in computational

complexity.

In Chapter 2, we provided an overview of CDMA and WCDMA. This chapter summa-

rized the features, which demonstrated CDMA’s superiority over other 2G cellular networks.

In addition, this chapter also introduced the requirements for 3G cellular systems, which

included wider bandwidth allocations, as well as new additional features.

User modeling helps compute the traﬃc density in the cellular network, which can be

used to optimize the placement of BSs and radio network controllers as well as to analyze

the performance of resource management algorithms towards meeting the ﬁnal goal: the

calculation and maximization of network capacity. In Chapter 3, we presented an analytical

model for approximating the user distributions in multi-cell WCDMA networks using 2-

dimensional Gaussian function by determining the means and the standard deviations of the

distributions for every cell. This allowed for the calculation of the intra-cell and inter-cell

interference, and the reverse-link capacity of the network. We analyzed three scenarios of

62

user distribution using 2-D Gaussian function: uniform distribution of users, users densely

clustered at the center of the cells, and users distributed at cells’ boundaries.

In Chapter 4, we maximized the capacity in multi-cell WCDMA networks. Optimizing

capacity was calculated with diﬀerent spreading factors of 256, 64, 16, and 4. Numerical

results showed that the SIR threshold for the received signals was decreased by 0.5 to 1.5

dB due to the imperfect power control. As expected, we can have many low rate voice users

or fewer data users as the data rate increases. The results also showed that the determined

parameters of the 2-dimensional Gaussian model matched well with the traditional method

for modeling user distribution.

In Chapter 5, we designed a call admission control algorithm for optimizing the through-

put in multi-cell WCDMA networks. Numerical results were analyzed for no, low, and high

mobility scenarios. The results showed that as the spreading factor increases, due to trunking

eﬃciency, the optimized throughput is better for all the three mobility model. Our method

of optimizing capacity and throughput is computationally eﬃcient and can be implemented

for large multi-cell WCDMA networks.

6.2 Future Research

We conclude by outlining possible directions for future research:

As shown in Table 4.2, the maximumtransfer data rate is only 480 Kbps when used with a

spreading factor of 4 in 5 MHz spectrum. To achieve the promised speed of 2 Mbps data rate

in 3G systems, 6 parallel codes with a spreading factor of 4 need to be used. Further research

can delve into how to use and interact between these parallel codes, and how the interference

level is generated and capacity is calculated in each cell in WCDMA networks with 2x60 MHz

spectrum (1710-1770 MHz and 2110-2170 MHz). The new IMT-2000 spectrum of 190 MHz

(2500-2690 MHz) arrangement is still under development.

63

BIBLIOGRAPHY

[1] 3GPP. 3GPP speciﬁcations - releases, phases, and stages.

http://www.3gpp.org/specs/releases.htm.

[2] R.G. Akl, M. Hegde, and M. Naraghi-Pour. Mobility-based CAC algorithm for arbitrary

traﬃc distribution in CDMA cellular systems. IEEE Trans. Veh. Technol., 54:639–651,

March 2005.

[3] R.G. Akl, M.V. Hegde, A. Chandra, and P.S. Min. CCAP: CDMA Capacity Allocation

and Planning. Technical report, Washington University, April 1998.

[4] R.G. Akl, M.V. Hegde, and P.S. Min. Eﬀects of call arrival rate and mobility on network

throughput in multi-cell CDMA. IEEE International Conf. on Commun., 3:1763–1767,

June 1999.

[5] R.G. Akl, M.V. Hegde, M. Naraghi-Pour, and P.S. Min. Multi-cell CDMA network

design. IEEE Trans. Veh. Technol., 50(3):711–722, May 2001.

[6] R.G. Akl and A. Parvez. Impact of interference model on capacity in CDMA cellular

networks. Proceedings of SCI 04: Communication and Network Systems, Technologies

and Applications, 3:404–408, July 2004.

[7] S. Anand, A. Chockalingam, and K. N. Sivarajan. Outage and capacity analysis of

cellular CDMA with admission control. IEEE Wireless Commun. and Networking Conf.,

2:908–912, March 2002.

[8] S. Ariyavisitakul and L.F. Chang. Signal and interference statistics of a CDMA system

with feedback power control. Global Telecommunications Conf., 2:1490–1495, December

1991.

64

[9] M. Asawa and W.E. Stark. Optimal scheduling of soft handoﬀs in DS/CDMA commu-

nication systems. Fourteenth Annual Joint Conf. of the IEEE Computer and Commun.

Societies, 1:105 – 112, April 1995.

[10] A. Baiocchi, F. Sestini, and F. Priscoli. Eﬀects of user mobility on the capacity of a

CDMA network. European Trans. on Telecommunications, 7(4):305–314, July-August

1996.

[11] P. Balaban and J. Salz. Optimum diversity combining and equalization in digital data

transmission with applications to cellular mobile radio I. Theoretical considerations.

IEEE Trans. Commun., 40:885–894, May 1992.

[12] P. Balaban and J. Salz. Optimum diversity combining and equalization in digital data

transmission with applications to cellular mobile radio II. Numerical results. IEEE

Trans. Commun., 40:895–907, May 1992.

[13] Mary Bellis. Selling the cell phones - history of cellular phones.

http://inventors.about.com/library/weekly/aa070899.htm.

[14] J.O. Carnero, K.I. Pedersen, and P.E. Mogensen. Capacity gain of an uplink-

synchronous WCDMA system under channelization code constraints. IEEE Veh. Tech-

nol. Conf., 53:982 – 991, July 2004.

[15] CDMA Development Group. CDG : Worldwide : 4Q 2004 CDMA Subscribers Statistics.

http://www.cdg.org/worldwide/cdma world subscriber.asp.

[16] CDMA Development Group. CDG : Worldwide : CDMA Worldwide.

http://www.cdg.org/worldwide/index.asp?h area=0.

[17] M. Chopra, K. Rohani, and J.D. Reed. Analysis of CDMA range extension due to soft

handoﬀ. IEEE Veh. Technol. Conf., 2:917 – 921, July 1995.

65

[18] G. Cimone, D.D. Weerakoon, and A.H. Aghvami. Performance evaluation of a two

layer hierarchical cellular system with variable mobility users using multiple class ap-

plications. IEEE Veh. Technol. Conf., 5:2835–2839, September 1999.

[19] T. Dohi, M. Sawahashi, and F. Adachi. Performance of SIR based power control in

the presence of non-uniform traﬃc distribution. IEEE International Conf. on Universal

Personal Commun., pages 334–338, November 1995.

[20] J.S. Evans and D. Everitt. On the teletraﬃc capacity of CDMA cellular networks. IEEE

Trans. Veh. Technol., 48(1):153–165, January 1999.

[21] Adachi F., Sawahashi M., and Okawa K. Tree-structured generation of orthogonal

spreading codes with diﬀerent lengths for forward link of DS-CDMA mobile radio.

Electr. Lett., 33:27–28, January 1997.

[22] E.D. Fitkov-Norris and A. Khanifar. Dynamic pricing in cellular networks, a mobility

model with a provider-oriented approach. 3G Mobile Commun. Technol., pages 63 – 67,

March 2001.

[23] G.J. Foschini and Z. Miljanic. A simple distributed autonomous power control algorithm

and its convergence. IEEE Trans. Veh. Technol., 42(4):641–646, November 1993.

[24] H. Fu and J.S. Thompson. Downlink capacity analysis in 3GPP WCDMA networks

system. 3G Mobile Commun. Technol., pages 534–538, May 2002.

[25] V.K. Garg. IS-95 CDMA and CDMA2000: Cellular/PCS Systems Implementation.

Prentice Hall communications engineering and emerging technologies series, 2000.

[26] K.S. Gilhousen, I.M. Jacobs, R. Padovani, A.J. Viterbi, L.A. Weaver, and C.E. Wheat-

ley. On the capacity of a cellular CDMA system. IEEE Trans. Veh. Technol., 40(2):303–

312, May 1991.

66

[27] S.V. Hanly. Capacity and power control in spread spectrum macrodiversity radio net-

works. IEEE Trans. Commun., 44(2):247–256, February 1996.

[28] S.V. Hanly. An algorithm for combined cell-site selection and power control to maximize

cellular spread spectrum capacity. IEEE J. Select. Areas Commun., 13(7):1332–1340,

September 1995.

[29] H. Holma and A. Toskala. WCDMA for UMTS: Radio Access for Third Generation

Mobile Communication. John Wiley & Sons, Ltd., 2002.

[30] infoSync World. infoSync World : 1.87 billion mobile users by 2007.

http://www.infosyncworld.com/news/n/5048.html.

[31] Y. Ishikawa and N. Umeda. Capacity design and performance of call admission control

in cellular CDMA systems. IEEE J. Select. Areas Commun., 15(8):1627–1635, October

1997.

[32] V.I. Istratescu. Fixed Point Theory : An Introduction. D. Reidel, 1981.

[33] S.H. Jeong and J.A. Copeland. Cell loss ratio and multiplexing gain of an ATM multi-

plexer for VBR voice sources. Local Computer Networks, pages 384–389, Oct 1998.

[34] Dan Jones. Air interfaces - what are they?

http://www.unstrung.com/document.asp?doc id=15195&page number=9.

[35] V.M. Jovanovic and D. Cuberovic. Theoretical and experimental analysis of the new

soft handoﬀ algorithm for CDMA systems. IEEE Veh. Technol. Conf., 3:1757 – 1762,

May 2000.

[36] F.P. Kelly. Blocking probabilities in large circuit switched networks. Advances in Applied

Probability, 18:473–505, 1986.

67

[37] F.P. Kelly. Routing in circuit-switched network: Optimization, shadow prices and de-

centralization. Advances in Applied Probability, 20:112–144, 1988.

[38] Kyoung II Kim. CDMA cellular engineering issues. IEEE Veh. Technol. Conf., 42:345–

350, August 1993.

[39] Z. Liu, Y. Wang, and D. Yang. Eﬀect of soft handoﬀ parameters and traﬃc loads on

soft handoﬀ ratio in CDMA systems. International Conf. on Communication Technology

Proceedings, 2:782 – 785, April 2003.

[40] David W. Matolak and Anjula Thakur. Outside cell interference dynamics in cellular

CDMA. Proceedings of the 35th Southeastern Symposium, pages 418–152, March 2003.

[41] D. Mitra, J.A. Morrison, and K.G Ramakrishnan. ATM network design and opti-

mization: a multirate loss network framework. IEEE/ACM Trans. on Networking,

4(4):531–543, August 1996.

[42] A.F. Naguib, A. Paulraj, and T. Kailath. Capacity improvement of base-station antenna

arrays cellular CDMA. Conf. Record of The Twenty-Seventh Asilomar Conf. on Signals,

Systems and Computers, 2:1437–1441, November 1993.

[43] S. Nguyen and R. Akl. Approximating user distributions in WCDMA networks using

2-D Gaussian. International Conf. on Comput., Commun., and Control Technol., July

2005, accepted.

[44] S. Nousiainen, K. Kordybach, and P. Kemppi. User distribution and mobility model

framework for cellular network simulations. IST Mobile & Wireless Telecommunications

Summit, pages 518–522, June 2002.

[45] T. Ojanpera, J. Skold, J. Castro, L. Girard, and A. Klein. Comparison of multiple

access schemes for UMTS. IEEE Veh. Technol. Conf., 2:490–494, May 1997.

68

[46] R. Padovani. Reverse link performance of IS-95 based cellular systems. IEEE Personal

Commun. Mag., 1(3):28–34, Third Quarter 1994.

[47] Bruce M. Penrod. A new class of precision UTC and fre-

quency reference using IS-95 CDMA base station transmissions.

http://www.endruntechnologies.com/pdf/PTTI2001 WhitePaper.pdf.

[48] Theodore S. Rappaport. Wireless Communications: Principles and Practice. Prentice

Hall, 2001.

[49] T.S. Rappaport. Wireless Communications Principles and Practice. Prentice Hall, 1996.

[50] K.V. Ravi. Comparison of multiple-accessing schemes for mobile communication sys-

tems. IEEE International Conf. on Personal Wireless Commun., pages 152–156, August

1994.

[51] C.U. Saraydar and A. Yener. Capacity enhancement for CDMA systems through adap-

tive cell sectorization. IEEE Wireless Commun. and Networking Conf., pages 1139–

1143, September 1999.

[52] S. Seo, T. Dohi, and F. Adachi. SIR-Based transmit power control of reverse link for

coherent DS-CDMA mobile radio. IEICE Trans. on Commun., E81-B(7):1508–1516,

July 1998.

[53] S. Shin, C. Cho, and D. Sung. Interference-based channel assignment for DS-CDMA

cellular systems. IEEE Trans. Veh. Technol., 48(1):233–239, January 1999.

[54] Y.M. Siu and K.K. Soo. CDMA mobile systems with tailor made power control to each

mobile station. 3G Mobile Commun. Technol., pages 46–50, March 2000.

69

[55] D. Staehle, K. Leibnitz, K. Heck, B. Schroder, A. Weller, and P. Tran-Gia. Approxi-

mating the othercell interference distribution in inhomogenous UMTS networks. IEEE

Veh. Technol. Conf., 4:1640–1644, May 2002.

[56] B. Suard, A.F. Naguib, G. Xu, and A. Paulraj. Performance of CDMA mobile commu-

nication systems using antenna arrays. IEEE International Conf. on Acoustics, Speech,

and Signal Processing, 4:153 – 166, April 1993.

[57] S.C. Swales, M.A. Beach, D.J. Edwards, and J.P. McGeehan. The performance en-

hancement of multibeam adaptive base-station antennas for cellular land mobile radio

systems. IEEE Trans. Veh. Technol., 39:56 – 67, February 1990.

[58] K. Tachikawa. W-CDMA: Mobile Communications System. John Wiley & Sons, Ltd.,

2002.

[59] K. Takeo and S. Sato. Evaluation of a CDMA cell design algorithm considering non-

uniformity of traﬃc and base station locations. IEICE Trans. Fundamentals, E81-

A(7):1367–1377, July 1998.

[60] K. Takeo and S. Sato. The proposal of CDMA cell design scheme considering change

in traﬃc distributions. IEEE International Symposium on Spread Spectrum Techniques

and Applications, 1:229–233, September 1998.

[61] Y. Takeuchi. Results of multiple cell simulations of a CDMA cellular system focusing on

SIR based power control. IEEE International Conf. on Commun., 1:63–67, June 1998.

[62] C. Tellez-Labao, J.M. Romero-Jerez, and A. Diaz-Estrella. Capacity analysis of SIR-

based power-controlled CDMA systems under multipath fading. IEEE International

Symposium on Personal, Indoor, and Mobile Radio Commun., 3:1237–1241, September

2002.

70

[63] TIA/EIA. Mobile station-base station compatibility standard for wideband spread spec-

trum cellular systems. Technical report, IS-95B, March 1999.

[64] TIA/EIA. Mobile station-base station compatibility standard for dual-mode wideband

spread spectrum cellular system. Technical report, IS-95A, May 1995.

[65] C. Vargas, M. Hegde, and M. Naraghi-Pour. Implied costs for multi-rate wireless net-

works. J.Wireless Networks, 10:323 337, May 2004.

[66] C. Vargas, M.V. Hegde, and M. Naraghi-Pour. Blocking eﬀects of mobility and reserva-

tions in wireless networks. IEEE International Conf. on Commun., 3:1612–1616, June

1998.

[67] C. Vargas, M.V. Hegde, and M. Naraghi-Pour. Implied costs in wireless networks. IEEE

Veh. Technol. Conf., 2:904–908, May 1998.

[68] C. Vargas-Rosales. Communication Network Design and Evaluation Using Shadow

Prices. PhD thesis, Louisiana State University, 1996.

[69] A.J. Viterbi. CDMA Principles of Spread Spectrum Communication. Addison-Wesley,

1995.

[70] A.J. Viterbi, A.M. Viterbi, K.S. Gilhousen, and E. Zehavi. Soft handoﬀ extends CDMA

cell coverage and increases reverse link capacity. IEEE J. Select. Areas Commun.,

12:1281–1288, October 1994.

[71] A.J. Viterbi, A.M. Viterbi, K.S. Gilhousen, and E. Zehavi. Soft handoﬀ extends CDMA

cell coverage and increases reverse link capacity. IEEE J. Select. Areas Commun.,

12(8):1281–1288, October 1994.

[72] A.M. Viterbi and A.J.Viterbi. Erlang capacity of a power controlled CDMA system.

IEEE J. Select. Areas Commun., 11(6):892–900, August 1993.

71

[73] J.H. Wen, J.S. Sheu, and J.L. Chen. An optimum downlink power control method for

CDMA cellular mobile systems. IEEE International Conf. on Commun., 6:1738–1742,

June 2001.

[74] D. Wong and T. J. Lim. Soft handoﬀs in CDMA mobile systems. IEEE Personal

Commun. Mag., 4:6–17, December 1997.

[75] B. Worley and F. Takawira. Handoﬀ schemes in CDMA cellular systems. Proceedings

of Commun. and Signal Processing 1998, pages 255 – 260, Sept. 1998.

[76] Q. Wu. Performance of optimum transmitter power control in cellular radio systems.

IEEE Trans. Veh. Technol., 48:571–575, March 1999.

[77] R.D. Yates. A framework for uplink power control in cellular radio systems. IEEE J.

Select. Areas Commun., 13(7):1341–1347, September 1995.

[78] Q. Zhang. UMTS air interface voice/data capacity-part 2: forward link analysis. IEEE

Veh. Technol. Conf., 4:2730 – 2734, May 2001.

[79] Q. Zhang and O. Yue. UMTS air interface voice/data capacity-part 1: reverse link

analysis. IEEE Veh. Technol. Conf., 4:2725 – 2729, May 2001.

72

CAPACITY AND THROUGHPUT OPTIMIZATION IN MULTI-CELL 3G WCDMA NETWORKS Son Nguyen, B.S.

Thesis Prepared for the Degree of MASTER OF SCIENCE

UNIVERSITY OF NORTH TEXAS August 2005

APPROVED: Robert Akl, Major Professor Robert Brazile, Committee Member and Graduate Coordinator Steve Tate, Committee Member Krishna Kavi, Chair of the Department of Computer Sciences Oscar N. Garcia, Dean of the College of Engineering Sandra L. Terrell, Dean of the Robert B. Toulouse School of Graduate Studies

ACKNOWLEDGMENTS First and foremost, I would like to give my heartfelt thanks to my advisor Dr. Robert Akl for taking on multitude of roles that provided guidance and direction. During this journey, I had sometimes felt I could not progress any further, but his whole-hearted devotion and enthusiasm not only kept me on track but also lightened up the way to the completion of this work. Furthermore, I am highly indebted to the members of my committee Dr. Robert Brazile and Dr. Steve Tate for their careful reading and suggestions. I am also very grateful to my parents for their unconditional love and many years of support. Above all, I would like to thank Khanh Ha Nguyen who has been an inspiration and a partner from the very beginning.

ii

CONTENTS

ACKNOWLEDGEMENTS

iii

LIST OF TABLES

vii

LIST OF FIGURES

ix

1 INTRODUCTION 1.1 1.2 1.3 CDMA History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 1 2 2

2 CDMA AND WCDMA OVERVIEW 2.1 Introduction to CDMA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.1 2.1.2 2.1.3 2.1.4 2.1.5 2.2 Power Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Frequency Reuse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Voice Activity Detection . . . . . . . . . . . . . . . . . . . . . . . . . Cell Sectoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Soft Handoﬀ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4 4 7 8 8 9 9 10

WCDMA Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3 USER AND INTERFERENCE MODELING USING 2-D GAUSSIAN FUNCTION 17 3.1 3.2 3.3 3.4 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 17 18 20 21

Related Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . User and Interference Model Numerical Results 3.4.1 . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Uniform Distribution of Users . . . . . . . . . . . . . . . . . . . . . .

iii

. . . . .2 5.8. . . . . . . . . . . . . . . . . . . . Network Throughput . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Calculation of N . . . . . . . . . . . . . . . . . . . . . . . . . . .8 Introduction . . . . . . . . . . . . . . . .6.2 4. .3 5. . . Spreading and Scrambling . . . . . . WCDMA Capacity with Imperfect Power Control . . .2 4. . . . WCDMA Capacity Optimization with SF of 4 . . . . . . . . . . . . . . . . . . . . . . . . .3.4 5. . . . . .
29 29 29 31 32 33 35 37 37 39 41 43
WCDMA Capacity Optimization with SF of 256 . . .
WCDMA Throughput Optimization with SF of 256 . .
5 WCDMA CALL ADMISSION CONTROL AND THROUGHPUT 5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7 . . . .5 4. . . . . . . . . . WCDMA Call Admission Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . WCDMA Capacity Optimization with SF of 64 . . . . . . . . .4 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5
Users Densely Clustered at the Center of the Cells
. . . . . . .3 3. .6. . . . . .1 5. . . . . . . . . . . . . . . Numerical Results 5. . . . . . . . . . .
iv
. . . . . . . . . .
Conclusions .
Conclusions . . . . . . . .3 4. . . . . . . . . . . . . . . . . . . . . . . . . .4 4. .6. . . . . .7 5. . . . .4. . .
4 WCDMA CAPACITY 4.5 5. . Related Work . . . . . . . . . . . . . . . . Numerical Results 4. . . . .
21 26 28
Users Distributed at Cells’ Boundaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6. . . . . . . . . . . . . . . . . . . . . WCDMA Capacity Optimization with SF of 16 . . . . . . . . . . . . .2 3. . . . . . . . .
46 46 46 47 49 50 51 52 52 54
Mobility Model . . . . . . . . . . . . . . . . . . . . Feasible States . . . . . . . . . . . .1 . . . .4. . . . . .6 5. . . . . . . .1 4. . . . . . . . . . . . . . . . Maximization of Throughput . . .3 4. . . . .6 Introduction . . . . . . . . . . . . . . WCDMA Capacity with Perfect Power Control . . .1 4. . . . . . . . . . . . . . . . . . . . . .

54 56 56 60
Conclusions . . . . . . . . . Future Research .2 5. . . . . . .8. . . . . . . . .2 Summary . . . . . . . . . . .4 5. . . . . . . . . . . . . . . . . . . . . . .3 5. . . . . . . . . . . . . WCDMA Throughput Optimization with SF of 16 . . . .8. . . .9
WCDMA Throughput Optimization with SF of 64 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8. . .5. .
62 62 63
v
. . . . . . . WCDMA Throughput Optimization with SF of 4 . . . .1 6. . . . .
6 CONCLUSIONS 6. . . .

. . . 3. . . . . . . . . . . . . . . . . . . . 3. . . . .5 42
dB. . . . . . . . . .1 3. . . .4 Capacity calculation for uniform user distribution with SF = 64 and
Eb Io
= 7. . .1 5. . . 4. . . . . . . . . µ1 . . . . . . . . . . . . . . . . . . . . .6. . σ2. . . .2 The maximum number of users in 27 cells of WCDMA network as the values of σ1 and σ2 are increased from 100 to 400 while µ1 = 0 and µ2 = 0. . . . . . . . . .3 Functionality of the channelization and scrambling codes. . . . . . . . .5 40
dB. . . . . . . . . . 4. . . . . .3 The values of σ1. . . . . The maximum number of users is 133. . . . . . . . Capacity calculation for uniform user distribution with SF = 256 and
Eb Io
13
22
24
27 34 36
= 38
7. . . . . . .5 dB. . . . . . . . . . . . . . . . .6 Capacity calculation for uniform user distribution with SF = 4 and
Eb Io
= 7. . . . This results in users distributed uniformly in all BSs. .5 44 53 54
dB. . . . . . . . . . . . . . .2 4. . . . . . . . 3. . . . . . . . . . . . . . .2 The low mobility characteristics and parameters. . . 5. . . . . . . . . . . . . This results in users densely clustered around the BSs. . . . . . . The high mobility characteristics and parameters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. . . Uplink DPDCH data rates. . . . . . 4.LIST OF TABLES 2. . . . . . .
vi
. . . . . . . . . . . .1 4. . . . . . . . . . . . . . . . . . . . . . . . . . and µ2 for the 2-D Gaussian approximation of users clustered at the boundaries of the cells as shown in Fig. . . . . . . . . The maximum number of users in every cell for the 27 cell WCDMA network (with σ1 and σ2 are increased from 5000 to 15000 while µ1 = 0 and µ2 = 0). . . .1 Main diﬀerences between WCDMA and IS-95 air interfaces .5 Capacity calculation for uniform user distribution with SF = 16 and
Eb Io
= 7. . . . .

5.3

Calculation of N for uniform user distribution with SF = 256 and blocking probability = 0.02. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

5.4

Calculation of N for uniform user distribution with SF = 64 and blocking probability = 0.02. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

5.5

Calculation of N for uniform user distribution with SF = 16 and blocking probability = 0.02. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

5.6

Calculation of N for uniform user distribution with SF = 4 and blocking probability = 0.02. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

vii

LIST OF FIGURES 2.1 2.2 2.3 2.4 Comparison between FDMA, TDMA, and CDMA. . . . . . . . . . . . . . . Frequency Hopping Spreading Spectrum. . . . . . . . . . . . . . . . . . . . . Time Hopping Spreading Spectrum. . . . . . . . . . . . . . . . . . . . . . . . 2 GHz band spectrum allocation in Europe, Japan, Korea, and US (MSS = Mobile Satellite Spectrum). . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5 3.1 3.2 Development to all-IP for 3G services. . . . . . . . . . . . . . . . . . . . . . Inter-cell interference on cell i from users in cell j . . . . . . . . . . . . . . . 2-D Gaussian approximation of users uniformly distributed in the cells. σ1 = σ2 = 12000, µ1 = µ2 = 0. The maximum number of users is 548. . . . . . . . 3.3 Simulated network capacity where users are uniformly distributed in the cells. The maximum number of users is 554. . . . . . . . . . . . . . . . . . . . . . 3.4 2-D Gaussian approximation of users densely clustered around the BSs. σ1 = σ2 = 100, µ1 = µ2 = 0. The maximum number of users is 1026. . . . . . . . 3.5 Simulated network capacity where users are densely clustered around the BSs causing the least amount of inter-cell interference. The maximum number of users is 1026 in the network. . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6 2-D Gaussian approximation of users clustered at the boundaries of the cells. The values of σ1, σ2 , µ1 , and µ2 may be diﬀerent in the diﬀerent cells and are given in Table 3.3. The maximum number of users is 133. . . . . . . . . . . . 3.7 Simulated network capacity where users are clustered at the boundaries of the cells causing the most amount of inter-cell interference. The maximum number of users is only 108 in the network. . . . . . . . . . . . . . . . . . . . 4.1 4.2 Generation of OVSF codes for diﬀerent Spreading Factors. . . . . . . . . . . Relationship between spreading and scrambling. . . . . . . . . . . . . . . . . 28 33 34 26 25 25 23 23 12 15 19 4 6 6

viii

4.3 4.4 4.5

12.2 Kbps Uplink Reference channel. . . . . . . . . . . . . . . . . . . . . . . 64 Kbps Uplink Reference channel. . . . . . . . . . . . . . . . . . . . . . . . Average number of slot per sector for perfect and imperfect power control analysis with a Spreading Factor of 256. . . . . . . . . . . . . . . . . . . . .

35 36

39

4.6

Average number of slot per sector for perfect and imperfect power control analysis with a Spreading Factor of 64. . . . . . . . . . . . . . . . . . . . . . 41

4.7

Average number of slot per sector for perfect and imperfect power control analysis with a Spreading Factor of 16. . . . . . . . . . . . . . . . . . . . . . 43

4.8

Average number of slot per sector for perfect and imperfect power control analysis with a Spreading Factor of 4. . . . . . . . . . . . . . . . . . . . . . . 45 56 57 57 60

5.1 5.2 5.3 5.4

Average throughput in each cell for SF = 256. . . . . . . . . . . . . . . . . . Average throughput in each cell for SF = 64. . . . . . . . . . . . . . . . . . . Average throughput in each cell for SF = 16. . . . . . . . . . . . . . . . . . . Average throughput in each cell for SF = 4. . . . . . . . . . . . . . . . . . .

ix

CHAPTER 1

INTRODUCTION 1.1 CDMA History The global mobile communications market has expanded very rapidly [13]. From analog phone systems in the 70’s and 80’s, cellular phone systems have progressed to digital cellular systems in their second generation (2G) with Time Division Multiple Access (TDMA), Frequency Division Multiple Access (FDMA), and Code Division Multiple Access (CDMA) technologies in the 90’s. Society has seen the introduction of data services from 2.5G with Short Message Service and now new third generation (3G) mobile phone systems are being introduced where cellular phones can access Internet services, retrieving text, pictures, video and other documents, with a promised delivery speed of up to 2 Mbps. Almost 1.52 billion people were using mobile phones by the end of 2004, and according to the new study by the Yankee Group, this number could reach 1.87 billion (27.4% total world population) by the end of 2007 [30]. CDMA is the fastest-growing digital wireless technology since its ﬁrst commercialization in 1994. 240 million subscribers already existed worldwide by the end of 2004 [15]. The major markets for CDMA are North America, Latin America, and Asia (particularly Japan and Korea). In total, CDMA was adopted by more than 100 operators across 76 countries around the globe [16]. According to [25, 50], CDMA technology can oﬀer about 7 to 10 times the capacity of analog technologies and up to 6 times the capacity of digital technologies such as TDMA. The advantages over TDMA and FDMA technologies, such as voice quality, system reliability, and handset battery life, have created a multitude of research on CDMA systems.

1

CDMA and WCDMA technologies are introduced. This chapter summarizes features as well as key points.1. WCDMA is the emerging technology for 3G mobile phone systems with diﬀerent data rates on demand for serving diﬀerent services for users. In addition. and the standard deviations of the 2-D Gaussian function in every cell. While CDMA is a 2G cellular system with a main focus on voice services. soft handoﬀ. • WCDMA call admission control and throughput: – Formulation and calculation of Call Admission Control (CAC) in WCDMA networks based on the QoS and Grade of Service (GoS) constraints. and cell sectoring. the 2-dimensional (2-D) Gaussian function is used to model user distribution in cellular networks. capacity in Wideband Code Division Multiple Access (WCDMA) cellular networks is calculated and optimized for diﬀerent data rate services.3 Organization In Chapter 2. user activity detection. the means. In addition.
1. which demonstrate CDMA’s superiority over other 2G cellular networks. • WCDMA capacity: – Formulation and calculation of the maximum capacity for diﬀerent services in WCDMA networks based on given quality of service (QoS) constraints. The objectives of this work are as follows: • User distribution modeling with 2-D Gaussian function: – Modeling user distribution by determining the user densities. this chapter 2
. This leads to the maximization of the throughput in WCDMA networks.2 Objectives In this work. These superior features include power control.

a call admission control algorithm is designed that maximizes the throughput in multi-cell WCDMA networks. as well as new additional features of 3G WCDMA cellular networks. Numerical results are presented for diﬀerent spreading factors and for several mobility scenarios. the formulation and calculation of the maximum capacity in WCDMA cellular networks is described. In Chapter 3. Our optimization with given quality of service constraints can ﬁnd the maximum number of simultaneous users for voice and data services in 3G networks. which summarize the contributions of this work. This method when used to calculate the average interference can more rapidly and eﬃciently compute the capacity in WCDMA networks when compared to other methods. the 2-D Gaussian function is used in modeling users in cellular networks. In Chapter 5. which are computationally intensive. Finally. In Chapter 4. like actual interference. in Chapter 6.also introduces the requirements for 3G cellular systems. which include wider bandwidth allocations. the conclusions are presented.
3
.

an analysis made by the Telecommunications Research and Action Center found CDMA outperformed other digital and analog technologies on every front. security. These advantages include increased capacity. Compared to existing technologies such as FDMA and TDMA. and soft handoﬀ mechanism. and reliability.CHAPTER 2
CDMA AND WCDMA OVERVIEW 2. According to [25]. which are assigned individually to each call.1 depicts the diﬀerences between the three technologies.1 Introduction to CDMA Code Division Multiple Access (CDMA) is a fairly new wireless communication technology.1: Comparison between FDMA. to carry multiple calls. 2. which was introduced by Qualcomm in the early 90’s. voice activity. power consumption. TDMA.
4
. immunity to multi-path fading. including signal quality. CDMA distinguishes diﬀerent calls by unique codes.
Figure 2. These gains and features yield better longevity of CDMA handset battery life and higher quality of voice signals. CDMA has many advantages over TDMA and FDMA technologies in cellular networks. and CDMA. which use diﬀerent frequency sub-bands and time slots respectively. Fig.

only one time slot is used to transmit the message.1. Direct Sequence Spread Spectrum modulates the data with a fast pseudo-random code sequence. With DS-CDMA. thus. Each user’s signal is distinguished by its special code. The essential idea behind SS is spreading the information signal over a wider bandwidth to make jamming and interception more diﬃcult. For each individual user. The ﬁrst type of spread spectrum is known as Frequency Hopping in which the signal is broadcasted over a seemingly random series of radio frequencies. and Direct Sequence (DS). the receiver. other signals from
5
. The only way to despread the signal to recover the message for a user is to use each user’s own special code. Time Hopping is the second type of SS where the transmission time is divided into equal amounts of time intervals called frames. which is assigned by the Base Station (BS). Time Hopping. into a wideband signal transmitted over the air medium. At this time. the signal is spread across a wider frequency band in direct proportion to the number of bits used. Each frame is divided into time slots. as shown in Fig. DS is the prefered spreading technique in CDMA and WCDMA and thus is the only one considered in this work.2.3. CDMA users share the same frequency spectrum. picks up the message. each bit in the original signal is modulated to multiple bits in the transmitted signal by using the spreading code. With the spreading code. The third spreading technique is Direct Sequence. With DS. hopping between frequencies in synchronization with the transmitter. each user’s signal is modulated with a high-rate special code (pseudorandom binary sequence). During each frame. hopping from frequency to frequency at a ﬁxed interval. By knowing the hopping sequence. as shown in Fig. Only the subscribed BS knows all its users’ special codes. Time Hopped systems assume that the sender and the receiver know the length of each time frame and the sequence of time slots in which the message will be modulated. 2. which is contained in the spreading code. CDMA signals appear to overlap in the time and frequency domain as was shown in Fig. 2.CDMA is based on Spread Spectrum (SS) communication. There are three types of SS: Frequency Hopping. a technique which was developed during World War II. 2.

2: Frequency Hopping Spreading Spectrum.
6
.3: Time Hopping Spreading Spectrum.
Figure 2.Figure 2.

23. the capacity of a CDMA network is limited by the amount of interference that is generated by all users in the network (unlike FDMA and TDMA capacities. In [73. the signal for each user is required to be above a given signal-tointerference ratio (SIR) threshold for it to be received and despread by the BS correctly. the authors propose an adaptive SIR based feedback power control. 27. which results in increasing longevity of the MS. Research on power control algorithms [19. power consumption can be signiﬁcantly reduced in the MS. In a CDMA network. which tries to solve the near-far problem while maintaining a low co-interference eﬀect by individually adjusting the SIR threshold control level for each mobile station. Power control aims to reduce interference by minimizing the eﬀects of the near-far problem (the received power at a BS from a mobile station (MS) near the cell boundary is less compared to a MS close to the BS). In [8]. and fading while keeping the received signal power. This control level is established by using Fuzzy Logic Control. with respect to its own radio link. In [54]. at the same level at the BS. 76] the authors deﬁne the problem of SIR balancing to be an eigenvalue problem in a link-gain matrix. Therefore. and try to ﬁnd the optimum downlink power control. 61. or the SIR. co-channel interference. 28. the authors investigate a feedback power control approach that allows power commands to be updated at a faster rate compared to the rate of multipath fading.1 Power Control Since the capacity of a CDMA network is interference-bound. the study of capacity characteristics focuses primarily on the methods of reducing interference. There are two main diﬀerent methods for managing power control: open-loop and closed-
7
. which are ﬁxed and primarily bandwidth limited). 52. Fast and precise power control is a key requirement for CDMA technology.
2. 77] has substantially investigated the near-far problem. In addition.diﬀerent users appear as noise and are represented as interference generated by the system.1.

the frequency reuse factor may be 4. Because a cellular network is given a pre-determined range of frequencies to exchange information between the BSs and MSs. In a hexagonal cellular network structure. is smaller. then the network can obtain a higher capacity. the total bandwidth is divided into a number of channels.1. where the MSs and BSs use the whole bandwidth to transfer and receive information.
2. reducing multiple access interference from neighboring cells results in a capacity gain. or 12. if the frequency reuse factor. Open-loop power control determines the transmit power such that the sum of transmit power and the received power is kept constant. in which one party signiﬁes the other party to increase or decrease its transmit power by a power step such that the target received
Eb Io
achieves a given threshold. 33]. The transmitted signal is attenuated with both fast (Rayleigh) and slow fading. Since CDMA systems use speech coding. 7. According to [33]. CDMA cellular systems typically use universal frequency reuse (or a frequency reuse factor of 1).BS and MS.
2.3 Voice Activity Detection In CDMA systems. This concept is called frequency reuse capability. the authors show that increasing the voice activity factor or
8
. diﬀerent cells in a cellular network can use the same frequencies if the signals from the same frequencies from one cell will not reach the BS of another cell. reducing the rate of the speech coder with voice activity detection along with variable rate data transmission could decrease the multiple access interference. In [24. Closed-loop power control involves with two parties . the voice activity factor for human speech averages about 42%.1. Thus. with precise power control.loop.2 Frequency Reuse In TDMA and FDMA. which the cellular network can employ. Each user is assigned a channel for an uplink (the frequency that a MS sends information to a BS) and a downlink (the frequency that a BS transmits the signal to a MS).

Cell Sectoring is a method that uses multiple directional antenna arrays to reduce the co-channel interference. Experimental data in [70] have shown that using soft handoﬀ increases cell coverage resulting in increased capacity
9
. On the uplink side.
2. as well as to transmit its signal to all the BSs. each individual MS can isolate and align both in time and in phase to reinforce the forward signals from diﬀerent BSs.
2.5 Soft Handoﬀ Soft handoﬀ (handover) is one of the most attractive features of CDMA technology. 75]. the authors in [11. 12. Depending on the relative received signal strength. By employing universal frequency reuse and a Rake receiver (see [58] pages 49-56). Soft handoﬀ is a technique that allows a MS in transition between two or more adjacent cells to transmit and receive the same signal from these BSs simultaneously. Recent improvements in soft handoﬀ algorithms are described in [9. the frequencies are permanently assigned to users and BSs as long as there is a communication between them. In FDMA and TDMA cellular systems.1. decisions are made as to when to enter soft handoﬀ and when to release the weaker BS.4 Cell Sectoring Due to increasing demand for cellular communication without corresponding increases in bandwidth allocation. 56]. the Mobile Switching Center must ﬁrst combine and resolve the signals from all the BSs and then determine which BS is receiving the stronger and better replica. According to [42.data activity factor could reduce the network capacity signiﬁcantly. 57] introduce Cell Sectoring (Spatial Processing) to improve spectrum utilization. Thus. resulting in increased cellular network capacity. the capacity in FDMA and TDMA systems is ﬁxed regardless whether the systems employ voice or data activity detection. both in receiving and transmitting.1. resolving angular positions of the mobiles by using antenna arrays at BSs. can lead to many-fold increases in system capacity.

which translates to range extension for CDMA cellular networks. In this work. average inter-cell interference is used to model and calculate capacity in WCDMA cellular networks. the authors compare the network capacity gain with diﬀerent parameter sets for new soft handoﬀ algorithm in IS-95A [64] and IS-95B [63]. 51. the authors analyze both reverse link and forward link capacity under the assumption of ideal power control and hard handoﬀ circumstances. These studies conclude that the capacity of the reverse link is lower than the capacity of the forward link. which includes person-to-person communication with high-quality images and video and high rate accessing
10
. 46. 59. The reverse link capacity limits the system capacity. accurate user modeling becomes essential for average interference calculation. however.
2. 53. Recent surveys [39. only the reverse link capacity is considered.6 dB. In addition. In [17].2 WCDMA Overview Third generation (3G) systems are designed for multimedia communication. Thus. 38. 40. the interference caused by diﬀerent users in the same cell is identical. Soft handoﬀ. 62] use average inter-cell interference instead of actual inter-cell interference for capacity analysis. 74] of soft handoﬀ show what technical issues need to be resolved. only a small diﬀerence exists between forward and reverse link capacity. however.6-3. 60.in CDMA networks. and is independent of their exact location within a cell. In this work. the shadow fading margin required by CDMA soft handoﬀ is less than TDMA/FDMA systems by 2. In [35]. requires complex design and implementation. Most of the models [7. 72]. 31. In an average inter-cell interference model. a comparison between hard handoﬀ (in TDMA/FDMA) and soft handoﬀ (in CDMA) shows that under a variety of conditions. 71. the surveys show what the beneﬁts and tradeoﬀs of using soft handoﬀ and discuss feasible parameter settings. 20. Extensive research has been done on calculating the reverse link capacity of single and multi-cell CDMA networks [10. In [26].

Within 3GPP. and CDMA2000. and CDMA2000 [34]. music. In the United States.. The process of developing third generation mobile systems started in a 1992 meeting of the World Administrative Radio Conference (WARC). Korea. Japan. not only for the cellphone. WCDMA speciﬁcations have been created in 3GPP (3rd Generation Partnership Project). these third generation systems are called International Mobile Telephony 2000 (IMT-2000). The two spectrums for IMT2000 TDD bands (1900-1920 MHz and 2020-2025 MHz) are available in Europe and Korea. W-CDMA.4 [29]. which is under the International Telecommunication Union (ITU).
11
. The spectrum allocation for IMT-2000 (or WARC-92) bands of 2x60MHz (1920-1980 MHz plus 2110-2170 MHz) are available in Europe. paid video games. at the time of the 1992 meeting. part of the IMT-2000 spectrum TDD is used by cordless telephone systems. Japan. third generation services can be implemented using the existing PCS spectrum with alternative technologies. In the United States. There are three major air interfaces proposed for 3G cellular systems: Enhanced Data Rate for GSM Evolution (EDGE). the conference identiﬁed that the frequencies around 2 GHz were available in most countries to be used by the future third generation mobile systems. infrastructure. but also for the applications and content providers carried by these networks (i. which is the joint standardization project of the bodies from Europe. the name WCDMA is used to cover both FDD and TDD operations. the US. Under ITU. WCDMA. and hardware manufactures and their operators. including EDGE. video. Korea. In Japan. and most Asian countries. and ring tones downloading).e. WCDMA is identiﬁed as UTRA (Universal Terrestrial Radio Access) FDD (Frequency Division Duplex) and TDD (Time Division Duplex).of information and services on public and private networks. and China. 2. no new spectrum had yet been made available for third generation systems. both terrestrial and satellite as shown in Fig. In the meeting. up to 2 Mbps. These new systems will create new business opportunities.

the following characteristics are the new requirements of 3G systems: • Data communication speeds up to 2 Mbps. Japan. including voice. and 806-960 MHz According to [29]. Korea. In the past. the following additional frequency bands were also introduced for IMT-2000: 1710-1885 MHz. • Mix of services with diﬀerent quality requirements on a single connection. and packet data. 2500-2690 MHz.Figure 2.
12
. 3G systems face the challenge of making data services wireless. The new IMT-2000 spectrum 190 MHz (2500-2690 MHz) arrangement is still under discussion. now. In addition. new third generation spectrums in the United States are expected to have 2x60 MHz (1710-1770 MHz and 2110-2170 MHz) assigned. These two spectrums can be eﬃciently used to carry 3rd generation services with WCDMA.4: 2 GHz band spectrum allocation in Europe. At the ITU-R WRC-2000 in May 2000. • Variable bit data rates on demand. video. 2G cellular mobile phone systems mainly concentrated on voice traﬃc. and US (MSS = Mobile Satellite Spectrum).

e.g. both uplink and downlink Not needed Yes.Table 2. web browsing results in more downlink than uplink traﬃc. • Coexistence with second generation systems and feasible solutions for inter-system handovers. • Quality of service (QoS) requirements from 10% frame error rate to 10−6 bit error rate. typically obtained via GPS Possible. • Higher spectrum eﬃciency. provides required quality of service
IS-95 1. • Coexistence of FDD and TDD modes. using FDD with a total
13
. IS-95 is a second generation cellular network. measurements with slotted mode Yes. downlink: slow power control Yes. but measurement method not speciﬁed Not needed for speech only networks
Packet data transmitted as short circuit switched calls
Downlink transmit diversity
Supported for improving downlink capacity
Not supported by the standard
• Meet delay requirement constraints for diﬀerent real-time services. The main diﬀerences between WCDMA and IS-95 are shown in Table 2.1: Main diﬀerences between WCDMA and IS-95 air interfaces
WCDMA Carrier Spacing Chip rate Power control frequency Base station synchronization Inter-frequency handovers Eﬃcient radio resource management algorithms Packet data Load-based packet scheduling 5 MHz 3.2288 Mcps Uplink: 800 Hz. which operates on the same frequency band as the Advanced Mobile Phone System (ﬁrst generation cellular network).25 MHz 1. which include coverage enhancements and load balancing. including delaysensitive real-time traﬃc (stock information) to best-eﬀort packet data (video or voice). • Support asymmetric uplink and downlink traﬃc.84 Mcps 1500 Hz..1 [29].

and email) were using circuit switched core networks. In March 2000. 3GPP Release ’99 kept voice and video services under circuit switched. 29]. while transforming SMS. The macro cell BSs are required to be located on masts or rooftops where GPS signals can be easily received. Inter-frequency handoﬀ is essential in WCDMA to maximize the use of several carriers
14
. thus. Also. WAP. and email to packet switched core networks. In 2G and 2. Web.5 [29]. IS-95 BSs are required to be synchronized and this can be done through the GPS system [47].bandwidth of 25 MHz in each direction [25]. while IS-95 uses fast power control only in the uplink. However. BSs in WCDMA are designed to operate under asynchronous mode. The asynchronous BSs have made the handoﬀ in WCDMA diﬀerent from IS-95. was developed [47]. all services (voice. Since GPS reception is diﬃcult without line-of-sight connection to the GPS satellites. where no synchronization from GPS is necessary. WAP. which improves the coverage of a network when cellular equipments are employed with Rake receivers [17]. MMS. These releases have led to WCDMA implementations to work with asynchronous BSs. The inclusion of fast power control at the MS side improves link performance and increases downlink capacity. The latest releases 5 and 6 of 3GPP have made all services use packet switched core networks [1. no synchronization from GPS is needed. SMS. it increases the complexity in the design of the MS. WCDMA uses a higher chip rate of 3. The IS-95 system was designed to target mainly macro cellular applications. and Streaming services were added to packet switched core networks. the deployment of indoor and micro cells was a challenge until recent technology. WCDMA changed the way communication is done in the core networks as shown in Fig.5G CDMA cellular systems. the higher chip rate in WCDMA enables higher bit rates while providing more multipath diversity.84 Mcps. Both WCDMA and IS-95 use direct sequence CDMA. 2. WCDMA uses fast closed-loop power control in both uplink and downlink. According to [29]. indirect GPS.

15
.Figure 2.5: Development to all-IP for 3G services.

16
. while in IS-95 inter-frequency measurements are not speciﬁed.per BS.

and an estimation on the amount of sites. With a given network setting (i. capacity calculation..1 Introduction Radio Network Planning is the problem of dimensioning.
17
. the problem of modeling user distribution and mobility has been an engaging research subject. User Modeling in WCDMA is a major process in the dimensioning activities. which can be used to optimize the placement of BSs and radio network controllers as well as to analyze the performance of resource management algorithms towards meeting the ﬁnal goal: the calculation and maximization of network capacity. the dimensioning activities determine the user distribution. Speciﬁcally. available spectrum. which is essential for the accurate calculation of interference and mobility. base station hardware. area type. quality of service. equipments at diﬀerent interfaces. and BS characteristics) and a path loss model.
3.CHAPTER 3
USER AND INTERFERENCE MODELING USING 2-D GAUSSIAN FUNCTION 3. blocking probability. radio network controllers. User Modeling helps compute the traﬃc density in the cellular network. which is a process of ﬁnding possible conﬁgurations and the amount of network equipment needed for an operator’s coverage.2 Related Work In the past decade. and radio propagation requirements. capacity. and core network elements. including circuit switched domain and packet switched (IP based) domain core network. dimensioning activities involve radio link budget and coverage analysis.e.

video. Recent research has looked at diﬀerent aspects that contribute to user distribution and mobility. This section will show that by using the 2D-Gaussian function at each BS. the authors’ model diﬀerent aspects of user distribution and mobility by taking two new concepts into account: user classes and street types. With the ability to load real map data into numerical simulations. In [22]. such as reduced quality video. which have larger coverage.3 User and Interference Model This study assumes that each user is always communicating and is power controlled by the base station (BS) that has the highest received power at the user. y) 18
. the software can calculate the network capacity and capture the movement of diﬀerent user classes in Helsinki area. including voice. the authors show that user distribution does drastically aﬀect the overall network capacity. and users at the cells’ boundaries. users densely clustered at the center of the cells. can be responsible for both new and handoﬀ requests but have limited application capabilities. In [6]. y) and rj (x. In [44].
3. This speciﬁc arrangement made the Macrocells encompass and lower the handoﬀ rate for faster-moving users.In [18]. Microcells accommodate both new and handoﬀ requests and can handle all types of high-rate connection requests. the authors use dynamic pricing to regulate the demand on wireless services. while Macrocells. Let ri (x. and data. user distribution and interference can be easily modeled for many diﬀerent scenarios. while the Microcells can be deployed in hot-spot areas to provide high-rate services to stationary or slow-moving users. including users uniformly distributed. as well as altering the mobility of users during peak time while also utilizing network resources during oﬀ-peak hours by lowering prices to the users. but most of it has focused on mobility and manually placing users in simulations. the authors propose a method for creating two layers of hierarchical cellular networks: Macrocell and Microcell to address high and low mobility of users.

(g)
(3.g e(γσs) = Aj
2
Cj
m rj (x. and w(x.1)
where γ = ln(10)/10. y).g be the average inter-cell interference that all users nj.2)
inter The inter-cell interference density Iji from cell j to BS i from all services G becomes inter Iji =
1 W 19
G g=1
Iji . m ri (x. σs is the standard deviation of the attenuation for the shadow fading.Figure 3.g using services g with activity factor vg and received signal Sg at BS j impose on BS i. m ri (x. y)
(3. y) w(x.1. 3. κji.3)
. y)
(3. y) w(x. Modifying the average inter-cell interference given by [5]. which BS j services as showed in Fig. respectively. Let Iji. y) is the user distribution density at (x. the equation becomes [43]
(g) Iji
= Sg vg nj. y) dA(x. y) dA(x.g be the per-user (with service g) relative inter-cell interference factor from cell j to BS i. This user is power controlled by BS j in the cell or region Cj with area Aj . Let κji.1: Inter-cell interference on cell i from users in cell j be the distance from a user to BS i and BS j. This study assumes that both large scale path loss and shadow fading are compensated by the perfect power control mechanism. m is the path loss exponent. y). y).g
e(γσs) Aj
2
Cj
m rj (x.

The path loss coeﬃcient m is 4. The simulator used for comparison is an extension of the software tools CDMA Capacity Allocation and Planning (CCAP) [3].g .4 Numerical Results The results shown are for a twenty-seven cell network topology used in [5.5)
where M is the total number of cells in the network.
3. The shadow fading standard deviation σs is 6 dB. G = 1.g κji. y) =
x−µ1 2 y−µ2 2 η −1( ) −1( ) e 2 σ1 e 2 σ2 . The COST-231 propagation model with a carrier frequency of 1800 MHz. is used to determine the coverage region. Eq. Louis for numerical analysis of optimization techniques developed 20
W R
is 21.g needs to be calculated only once. These results are compared with simulations to determine the value of η experimentally.
j=1..
. The activity factor. is 0. 6].g .4)
Thus. (3. written in MATLAB. an approximation can be found for a wide range of user distributions ranging from uniform to hot-spot clusters.1 dB. The processing gain v.5 meters. The user distribution is modeled with a 2-dimensional Gaussian function as follows w(x. the total inter-cell interference density Iiinter from all other cells to BS i is Iiinter = 1 W
M G
Sg vg nj.
g=1
(3.j=i g=1
(3. i.6)
where η is a user density normalizing parameter. average base station height of 30 meters and average mobile height of 1. then. 2πσ1 σ2
(3. was developed at Washington University in St. If the user distribution density can be approximated.375. κji.e.g κji. CCAP. By specifying the means µ1 and µ2 and the standard deviations σ1 and σ2 of the distribution for every cell. We assume only one service.where W is the bandwidth of the system.3) can be rewritten as
inter Iji =
1 W
G
Sg vg nj.

5. users clustered at the center of the cells.1 shows the maximum number of users in every cell for the 27 cell WCDMA network. the users cause the least amount of interference to the network by reducing the power gain required to maintain a desired signal-to-noise ratio. almost uniformly in the cells.1 Uniform Distribution of Users The network with diﬀerent values of σ1 and σ2 has been analyzed.2 Users Densely Clustered at the Center of the Cells Table 3. which also yields a total number of users equal to 1026.in [5] to compute the capacity of CDMA networks.4. σ2 result in users spread out.4. while µ1 = 0 and µ2 = 0.
21
. This compares well with simulation results presented in Fig. The results are veriﬁed with [6]. The following models show that by using 2-D Gaussian distribution. Fig. This compares exactly with simulation results presented in Fig.6).2 shows the maximum number of users in every cell for the 27-cell WCDMA network as the values of σ1 and σ2 are increased from 100 to 400 while µ1 = 0 and µ2 = 0. This results in users densely clustered around the BSs. The maximum number of users is 1026.
3.3. as the values of σ1 and σ2 are increased from 5000 to 15000. Fig. Table 3. and users at the cells’ boundaries. 3. while keeping µ1 and µ2 equal to zero in (3.
3. These increments of σ1 . This study extends CCAP for WCDMA networks and uses the 2-dimensional Gaussian function for w(x. many diﬀerent scenarios can be modeled. where actual distances were used to simulate real-time users entering the network for the calculation of interference. y). which yields a total number of users equal to 554 when they are placed uniformly in the cells. The total number of users is 548. 3. 3.4 shows the 2-D Gaussian approximation with σ1 = σ2 = 100. 3.2 shows the 2-D Gaussian approximation of users uniformly distributed in the cells with σ1 = σ2 = 12000. In this conﬁguration. including users uniformly distributed.

This results in users distributed uniformly in all BSs.1: The maximum number of users in every cell for the 27 cell WCDMA network (with σ1 and σ2 are increased from 5000 to 15000 while µ1 = 0 and µ2 = 0).Table 3. σ2 Cell1 Cell2 Cell3 Cell4 Cell5 Cell6 Cell7 Cell8 Cell9 Cell10 Cell11 Cell12 Cell13 Cell14 Cell15 Cell16 Cell17 Cell18 Cell19 Cell20 Cell21 Cell22 Cell23 Cell24 Cell25 Cell26 Cell27 Total Users 5000 18 18 18 18 18 18 18 18 18 22 22 22 17 18 18 22 22 22 18 25 25 25 25 25 25 25 25 565 7000 18 18 18 18 18 18 18 18 17 21 22 21 17 18 17 21 22 21 17 25 25 25 25 25 25 25 25 558 10000 18 18 18 18 18 18 18 18 17 21 22 21 17 18 17 21 21 21 17 25 24 24 25 25 24 24 25 553 12000 18 18 17 17 18 17 17 18 17 21 21 21 17 18 17 21 21 21 17 25 24 24 25 25 24 24 25 548 15000 18 18 17 17 18 17 17 18 17 21 21 21 17 18 17 21 21 21 17 25 24 24 25 25 24 24 25 548 Capacity from [5] 18 18 17 17 18 17 17 18 17 21 21 21 17 18 17 21 21 21 17 25 24 24 25 25 24 24 25 548
22
.
σ = σ1 .

2: 2-D Gaussian approximation of users uniformly distributed in the cells. µ1 = µ2 = 0.
Figure 3. The maximum number of users is 548.3: Simulated network capacity where users are uniformly distributed in the cells.
23
. σ1 = σ2 = 12000.Figure 3. The maximum number of users is 554.

σ2 Cell1 Cell2 Cell3 Cell4 Cell5 Cell6 Cell7 Cell8 Cell9 Cell10 Cell11 Cell12 Cell13 Cell14 Cell15 Cell16 Cell17 Cell18 Cell19 Cell20 Cell21 Cell22 Cell23 Cell24 Cell25 Cell26 Cell27 Total Users σ = 100 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 1026 σ = 200 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 38 1026 σ = 300 37 37 37 37 37 37 37 37 37 37 37 37 37 37 37 37 37 37 37 37 38 38 38 38 37 37 37 1003 σ = 400 34 34 35 35 34 35 35 35 35 36 36 36 35 35 35 35 35 35 35 36 36 37 36 36 36 36 36 954
24
.2: The maximum number of users in 27 cells of WCDMA network as the values of σ1 and σ2 are increased from 100 to 400 while µ1 = 0 and µ2 = 0.
σ = σ1 .Table 3. This results in users densely clustered around the BSs.

µ1 = µ2 = 0.5: Simulated network capacity where users are densely clustered around the BSs causing the least amount of inter-cell interference.Figure 3. The maximum number of users is 1026. σ1 = σ2 = 100.
25
.
Figure 3.4: 2-D Gaussian approximation of users densely clustered around the BSs. The maximum number of users is 1026 in the network.

3 Users Distributed at Cells’ Boundaries Fig. The values of σ1 . The pattern seen in Fig. These results are close to what was attained through simulation. σ2.
26
. 3. The maximum number of users is 133. 3.6 shows the 2-D Gaussian approximation of users clustered at the boundaries of the cells. µ1 . The maximum network capacity was purposely decreased by the simulator by placing the users so they caused the maximum interference to the network.Figure 3. The values of σ1. σ2.3.4. The maximum number of users is 133.3. 3. and µ2 may be diﬀerent in the diﬀerent cells and are given in Table 3. and µ2 may be diﬀerent in the diﬀerent cells and are given in Table 3.6: 2-D Gaussian approximation of users clustered at the boundaries of the cells. The placement at extremities would require users to increase their power gain causing much more interference to other users.7 shows that the simulator placed the users at the extreme corners of their respective cells. The simulation yielded a total capacity of 108 users. µ1 . with only 4 users in each cell.

Table 3. µ1 . σ2 . The maximum number of users is 133.
µ1 Cell1 Cell2 Cell3 Cell4 Cell5 Cell6 Cell7 Cell8 Cell9 Cell10 Cell11 Cell12 Cell13 Cell14 Cell15 Cell16 Cell17 Cell18 Cell19 Cell20 Cell21 Cell22 Cell23 Cell24 Cell25 Cell26 Cell27 -1400 -1400 -1400 0 0 1300 -1400 -1300 0 0 0 -1400 0 1300 1300 -1350 -1400 0 -1400 -1400 -1350 0 1350 1400 0 0 -1350 σ1 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 µ2 -900 800 800 -1700 -1600 -800 900 900 1500 1600 1550 900 1500 900 -800 -850 -900 -1600 -800 -800 800 1600 800 -800 -1700 -1600 -850 σ2 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300
27
. 3.3: The values of σ1.6. and µ2 for the 2-D Gaussian approximation of users clustered at the boundaries of the cells as shown in Fig.

5 Conclusions An analytical model has been presented for approximating the user distributions in multi-cell WCDMA networks using 2-dimensional Gaussian function by determining the means and the standard deviations of the distributions for every cell. This allowed for the calculation of the inter-cell interference and the reverse-link capacity of the network. The maximum number of users is only 108 in the network. The model compares well with simulation results and is fast and accurate enough to be used eﬃciently in the planning process of large WCDMA networks.Figure 3.7: Simulated network capacity where users are clustered at the boundaries of the cells causing the most amount of inter-cell interference.
28
. 3.

CDMA technology can oﬀer about 7 to 10 times the capacity of analog technologies and up to 6 times the capacity of digital technologies such as TDMA. and handset battery life compared to TDMA and FDMA technologies.
4. the authors present a method to calculate the WCDMA reverse link Erlang capacity based on the Lost Call Held (LCH) model as described in [69]. system reliability. 50]. This algorithm
29
. In [79]. WCDMA. which found that WCDMA was well suited for supporting variable bit rate services. According to [25. the new 3G systems are required to have additional support on a variety of data-rate services using multiple access techniques. several research on WCDMA capacity has been considered. In total. CDMA is the fastest-growing digital wireless technology since its ﬁrst commercialization in 1994. The major markets for CDMA are North America. Latin America.1 Introduction 3G cellular systems are identiﬁed as International Mobile Telecommunications-2000 under International Telecommunication Union and as Universal Mobile Telecommunications Systems (UMTS) by European Telecommunications Standards Institute. and Asia (particularly Japan and Korea). 45]. Besides voice capability in 2G. CDMA has been adopted by more than 100 operators across 76 countries around the globe [16]. the next generation of CDMA. With its tremendous advantages such as voice quality. is the best candidate for 3G cellular systems [29.CHAPTER 4
WCDMA CAPACITY 4.2 Related Work Since the ﬁrst comparisons of multiple access schemes for UMTS [45].

the capacity is calculated at the BS. the potential capacity gain is about 35. In this work. signals transmitted from diﬀerent MSs are time aligned at the BS. But the implementation is fairly complex while the potential capacity gain is not very high. Synchronous WCDMA looks at time synchronization for signal transmission between the BS and MS to improve network capacity. In [14]. The access in the calculation of forward link capacity is one-to-many rather than many-toone as in the reverse link. An alternate approach. MSs in the same cell share the same scrambling code. The transmission power from the BS is provided to each user based on each user’s relative need. In this research.6% where soft handoﬀ is enabled. the capacity gain in a more realistic scenario is reduced to 9. i.8% in a multicell scenario with inﬁnite number of channelization codes per cell and no soft handoﬀ capability between MSs and BSs. The same LCH model was also used in [78] to calculate the forward link capacity of UMTS/WCDMA systems based on the system outage condition. The performance of an uplink-synchronous WCDMA is analyzed in [14]. The goal of this uplink-synchronous method in WCDMA is to reduce intra-cell interference. the authors provide capacity calculation results and performance evaluation through simulation.. Scrambling codes are unique for each cell.e. the authors derive a closed form expression of Erlang capacity for a single type of traﬃc loading and compare analytical results with simulations results. while diﬀerent orthogonal channelization codes are derived from the set of Walsh codes.calculates the occupancy distribution and capacity of UMTS/WCDMA systems based on a system outage condition. has been considered.
30
. we will calculate the maximum reverse link capacity in UMTS/WCDMA systems for a given quality of service requirements. In this research. However. In the forward link. because many users share the BS transmission power. where MSs are synchronized on the uplink.

g − vg
(4.5)
The capacity in a WCDMA network is deﬁned as the maximum number of simultaneous users (n1. Substituting (3.M.g − vg ≤ cef f .. 55] Eb I0 =
i. for i=1.3 WCDMA Capacity with Perfect Power Control In WCDMA.g . (4. .g in BS i for a given service g needs to meet the following inequality constraint τg ≤ N0 +
∗ Sg Rg ∗ Sg W
G g=1
M
G
.2) into (4....4)
where cef f =
(g)
W 1 Rg − ∗ .3) can be rewritten as
G M G
ni. G that satisfy (4.. which the user can
transmit.
(4. with perfect power control (PPC) between BSs and MSs. Thus Iiown is given by Iiown = 1 W
G
Sg vg ni. the number of users ni..g vg κji.1)
where N0 is the thermal noise density. nj. Iiinter was calculated in section 3..g vg +
j=1.M.5) and (4.3.g vg +
g=1 j=1.j=i g=1
nj.3)
ni.g Sg Rg
N0 + Iiinter + Iiown − Sg vg
..
(g)
(4. we have for every cell i in the WCDMA network. .
31
.g .
g=1
(4. and Rg is the bit rate for service g.. nM. and Sg be the maximum signal power.g vg κji.. .. . Rg τg Sg /N0
(4.j=i g=1
After rearranging terms.1). the energy per bit to total interference density at BS i for a service g is given by [43. Iiown is the total intra-cell interference density caused by all users in cell i.g) for all services g = 1.. for i=1.4).g .4. n2.2)
Let τg be the minimum signal-to-noise ratio. which must received at a BS to decode the
∗ signal of a user with service g.

transmitted signals between BSs and MSs are subject to multipath propagation conditions. inverting (4.g I0
i.g I0 where β = ln(10)/10.11)
32
. I0
(4.g (βσc )2 e 2 . in the imperfect power control (IPC) case.
(4.g = median I0 for every user with service g in cell i.g ]
= eβmc e
(βσc )2 2
.g
signals vary according to a log-normal distribution with a standard deviation on the order of 1.g in each cell i for every user with service g needs to be replaced by the variable (Eb )i.
(4.g = 10log10
i. We deﬁne xi. Hence.g
Eb I0
. we have E 10(xi.
which is log-normally distributed. the expected value becomes E (Eb )o.g
I0
.g /10 = eβxi.6)
to be a normally distributed random variable with mean mc and standard deviation σc .3.g
=
(Eb )o.g (Eb )o.g i. then taking the average value (the expected value).g I0
i.g is Gaussian
with mean mc and standard deviation σc . we have (Eb )o. Thus.6).
i.
(4.g I0
= E
(Eb )o.g
= 10xi. Thus.g . According to [69].g /10) We can choose
(Eb )o.10)
=
(Eb )i.5 to 2. assumes perfect power control between the BSs and MSs.g .g
(4. which was formulated in section 4.9)
which makes (Eb )o.7)
using the fact that xi.8)
such that (Eb )o.g = eβmc = 10mc /10. I0 (4.g E[ I0
i.g (Eb )o. which make the received
Eb Io i.5 dB [69]. However.g =
i. the constant value of (Eb )i.4.4 WCDMA Capacity with Imperfect Power Control The calculation of WCDMA network capacity. by evaluating the nth moment of
i.

Table 4.1: Generation of OVSF codes for diﬀerent Spreading Factors. Scrambling is used on top of spreading.
4.1 describes the diﬀerent functionality of the channelization and the scrambling codes. Fig. Fig. which were originally introduced in [21].11). 4. cef f
(g)
IP C
becomes cef f / e
(g)
(βσc )2 2
.1 depicts the generation of diﬀerent OVSF codes for diﬀerent SF values. As a result of (4. were used to be channelization codes for UMTS. The use of OVSF codes allows the orthogonality and spreading factor (SF) to be changed between diﬀerent spreading codes of diﬀerent lengths.e.5 Spreading and Scrambling Communication from a single source is separated by channelization codes. The Orthogonal Variable Spreading Factor (OVSF) codes.. thus it only makes the signals from diﬀerent sources distinguishable from each other. The data signal after spreading is then scrambled with a scrambling codes to separate MSs and BSs from each other.2 depicts the relationship between the spreading and scrambling process.
33
.Figure 4. i. the dedicated physical channel in the uplink and the downlink connections within one sector from one MS. 4.

1: Functionality of the channelization and scrambling codes. Length Uplink: 4-256 chips same as SF Downlink 4-512 chips same as SF Number of codes Number of codes under one scrambling code = spreading factor Code family Orthogonal Variable Spreading Factor Uplink: 10 ms = 38400 chips Downlink: 10 ms = 38400 chips Uplink: Several millions Downlink: 512 Long 10 ms code: Gold Code Short code: Extended S(2) code family Spreading Yes. does not aﬀect transmission bandwidth Scrambling code Uplink: Separation of MSs Downlink: Separation of sectors (cells)
34
.
Table 4.
Channelization code Usage Uplink: Separation of physical data (DPDCH) and control channels (DPCCH) from same MS Downlink: Separation of downlink connections to diﬀerent MSs within one cell.2: Relationship between spreading and scrambling.Figure 4. increases transmission bandwidth No.

Fig.2 Kbps.06 dB. average base station height of 30 meters and average mobile height of 1.4 shows the DPDCH data rate requirement for 64 Kbps data user. The typical required data rate or Dedicated Traﬃc Channel (DTCH) for a voice user is 12.3: 12. 4.Figure 4. Fig. and Radio Frame Alignment. and the processes of Channel Coding. 12. which is the actual transmitted data rate.04 dB. 18.
1 2
rate coding
4. However. 43]. The shadow fading standard deviation σs is 6 dB.6 Numerical Results The results shown are for a twenty-seven cell network topology used in [2.2 Kbps Uplink Reference channel. 4. Rate Matching. The processing gain
W Rg
is 6.02 dB. Table 4.2 shows the approximation of the maximum user data rate with for diﬀerent values of DPDCH. The COST-231 propagation model with a carrier frequency of 1800 MHz.5 meters. the Dedicated Physical Data Channel (DPDCH). 5. 6.08 dB for Spreading Factor
35
. The path loss coeﬃcient m is 4.3 depicts the process of creating the actual transmitted signal for a voice user. is dramatically increased due to the incorporated Dedicated Control Channel (DCCH) information. and 24. is used to determine the coverage region.

)
7.
Table 4.Figure 4.2: Uplink DPDCH data rates.8 Mbps
36
.4: 64 Kbps Uplink Reference channel.5 Kbps 15 Kbps 30 Kbps 60 Kbps 120 Kbps 240 Kbps 480 Kbps 2. with 6 parallel codes DPDCH channel bit rate (Kbps) 15 30 60 120 240 480 960 5740 Maximum user data rate with
1 2
rate coding (approx.
DPDCH Spreading Factor 256 128 64 32 16 8 4 4.

whose parameters were determined in Chap. This study extends the software tools (CCAP) [3] to analyze diﬀerent SF in WCDMA networks.5 dB. was developed at Washington University in St. 4. Table
37
. we set SF to 256. The network with diﬀerent values for values of 4. 64. Louis for numerical analysis of optimization techniques developed in [5] to compute the capacity of CDMA networks.5 also shows that the traditional uniform user distribution modeling matches well with the 2-D Gaussian model.5 dB
while the standard deviation of the imperfect power control is increased from 0 to 2.3. is 0. 64. v.5 dB to 1. Table 4. we have to decrease the SIR threshold by 0. which is used to carry data for the control channel. the number of slots per sector decreases by almost a factor of 4 compared to SF equal 256 (from 60.3 shows the number of slots per cell with omni-directional antenna for the scenario with
Eb I0
= 7. The WCDMA network with 27 omni-directional antenna cells (1 sector per cell) was analyzed for evaluation of capacity using user modeling with the 2-D Gaussian function. Because of IPC.equal to 4.5 dB in PPC).6.
Eb I0
was analyzed for diﬀerent SF
4. 4. and 256.
4. to get the same average number of slots per sector as PPC. 4.375. CCAP.56 slots when
Eb Io
= 7. As a result of lowering the SF to 64. which is used for voice communication as shown in Fig. and traditional methods of modeling uniform user distribution. written in MATLAB. 16.5 shows the optimized average number of slots per sector for the 27 cells WCDMA network as the
Eb I0
is increased from 5 dB to 10 dB and the standard deviation of imperfect
power control is increased from 0 to 2. Fig. we set SF to 64.5 dB. The activity factor. 16.2 WCDMA Capacity Optimization with SF of 64 Next. and 256.5 dB. 3. respectively.58 to 15.1 WCDMA Capacity Optimization with SF of 256 First.6. Fig.

80 47.79 51.04 36.49 73.62 51.29 62.3: Capacity calculation for uniform user distribution with SF = 256 and dB.77 52.19 36.80 45.87 37.64 51.5 79.40 71.87 37.83 71.24 55.20 38.92 46.64 52.74 45.08 62.76 51.20 38.46 43.64 Uniform User Distribution 0.79 51.27 62.74 52.51 50.42 53.20 37.00 36.01 50.88 51.65 71.41 43.54 60.84 53.88 45.38 45.89 1178.55 45.07 46.71 63.04 36.85 65.55 38.45 45.95 51.94 46.59 45.43 71.5 97.55 45.78 51.71 50.73 62.26 73.97 47.78 52.65 73.54 56.66 63.07 47.35 37.78 51.08 65.65 2.66 65.85 51.19 36.22 63.86 44.38 45.90 52.94 46.05 55.58 1.87 55.40 1635.25 63.22 55.0 110.92 46.38 52. σ: cef f Cell1 Cell2 Cell3 Cell4 Cell5 Cell6 Cell7 Cell8 Cell9 Cell10 Cell11 Cell12 Cell13 Cell14 Cell15 Cell16 Cell17 Cell18 Cell19 Cell20 Cell21 Cell22 Cell23 Cell24 Cell25 Cell26 Cell27 Network Capacity Average Capacity per Sector 2-D Gaussian function 0.51 63.10 45.49 52.54 44.08 62.73 62.86 53.10 63.73 50.83 63.40 38.87 46.82 73.74 63.08 55.08 47.10 1448.62 52.87 55.95 38.10 63.74 73.04 45.20 36.43 73.73 53.51 50.20 52.20 37.95 51.07 47.72 56.5 97.73 45.5
38
.62 52.24 65.05 55.00 50.26 1630.0 110.09 2.36 37.66 65.75 37.84 53.87 65.22 1453.54 56.75 37.00 36.84 63.64 51.60 45.Table 4.10 65.88 47.
User Modeling with Imperfect Power Control.19 36.5 79.45 45.08 55.74 50.65 71.40 37.84 51.75 51.38 38.35 38.26 71.07 46.20 52.75 56.64 53.29 54.26 71.86 44.54 44.56 52.88 46.04 45.37 1.91 51.80 45.20 45.49 62.66 63.81 65.65 73.06 60.84 71.88 45.78 1174.86 73.09 38.35 38.67 53.74 50.50
Eb Io
= 7.62 51.49 62.18 45.10 45.51 63.

SF = 256 10 PPC (Gaussian). the SIR threshold has to be decreased by 0.4.5 dB IPC (Gaussian).5 dB PPC (Uniform).5 dB to 1.5: Average number of slot per sector for perfect and imperfect power control analysis with a Spreading Factor of 256.
4.56 to 4.3 WCDMA Capacity Optimization with SF of 16 Next.5 dB IPC (Uniform).
.6 shows the optimized average number of slots per sector for the 27 cells WCDMA network as the
Eb I0
is increased from 5 dB to 10 dB and the standard deviation of imperfect power control is increased from 0 to 2. σ = 1.6 also shows that the traditional uniform user distribution modeling matches well with the 2-D Gaussian model.30 slots when 39
Eb Io
= 7.5 dB. σ = 0 dB IPC (Uniform). Fig. the number of slots per sector decreases by almost a factor of 4 compared to SF equal 64 (from 15. σ = 2. we set SF to 16. Fig. to get the same amount of the average number of slots per sector as PPC.4 shows the number of slots per cell for the scenario with
Eb I0
= 7.5 dB. σ = 1. which is used for 64 Kbps data communication as shown in Fig.5 dB in PPC).5 dB.5 dB while the standard
deviation of the imperfect power control is increased from 0 to 2.5 dB
9
SIR threshhold in (dB)
8
7
6
5 20 40 60 80 100 Average number of slots per sector 120
Figure 4. σ = 2. 4. 4. 4. Because of IPC. As a result of lowering the SF to 16. σ = 0 dB IPC (Gaussian). 4.6.

56 14.11 16.20 16.56 12.20 16.40 18.82 18.70 13.97 18.81 374.44 13.39 13.69 10.48 11.60 9.87 2.69 9.86 13.81 16.85 10.48 11.72 9.30 13.79 11.56 14.73 13.49 14.
40
.72 13.23 12.14 11.44 16.82 2.58 13.61 13.83 11.06 12.06 9.06 16.08 11.87 11.08 11.44 11.41 16.72 305.25 16.62 13.43 13.65 9.86 9.81 16.45 16.35 11.06 9.4: Capacity calculation for uniform user distribution with SF = 64 and
User Modeling with Imperfect Power Control.82 11.77 11.65 9.50 13.51 1.86 18.78 16.43 13.70 13.32 11.82 13.03 16.09 9.87 11.36 14.53 13.10 12.32 13.30 13.31 14.39 13.26 13.45 18.82 18.82 418.48 9.88 10.07 15.32 13.65 9.60 9.72 11.5 20.78 373. σ: cef f Cell1 Cell2 Cell3 Cell4 Cell5 Cell6 Cell7 Cell8 Cell9 Cell10 Cell11 Cell12 Cell13 Cell14 Cell15 Cell16 Cell17 Cell18 Cell19 Cell20 Cell21 Cell22 Cell23 Cell24 Cell25 Cell26 Cell27 Network Capacity Average Capacity per Sector 2-D Gaussian function 0.43 13.83 11.11 13.81 16.29
Eb Io
= 7.0 28.46 18.69 11.85 18.43 13.26 13.44 14.73 9.45 18.31 14.75 11.32 11.87 12.Table 4.5 dB.44 9.06 16.45 18.5 25.69 9.78 16.67 15.69 9.45 16.14 11.49 13.69 9.97 16.82 13.44 14.26 13.70 13.70 304.97 16.69 11.5 25.32 13.03 13.62 12.45 16.44 11.97 13.36 11.37 14.32 13.62 14.35 11.06 12.41 16.65 10.56 1.03 18.91 9.82 18.32 Uniform User Distribution 0.11 13.31 11.49 14.39 13.26 13.5 20.56 16.86 18.62 16.82 12.41 16.39 13.26 16.86 13.40 18.87 12.11 16.36 11.79 11.86 9.03 16.82 11.69 9.78 16.82 12.09 9.85 420.0 28.23 12.40 18.73 11.62 14.40 18.25 13.41 16.54 12.54 12.31 11.73 13.06 12.91 9.

4. we set SF to 4.6.5 dB. 4. σ = 1. to get the same average number of slots per sector as PPC.5 dB in PPC. σ = 2. 4. σ = 0 dB IPC (Gaussian).SF = 64 10 PPC (Gaussian).5 dB PPC (Uniform). σ = 0 dB IPC (Uniform).5 shows the number of slots per cell for the scenario with
Eb I0
= 7. σ = 1. Fig.4 WCDMA Capacity Optimization with SF of 4 Next.5 dB IPC (Uniform). which is used for 256 Kbps data communication between BSs and MSs.49 while keeping
Eb Io
= 7.5 dB to 1.6 shows the number of slots per cell for the 41
.5 dB IPC (Gaussian). As a result of lowering the SF to 4.5 dB. Table 4.5 dB. we have to increase the SIR threshold by 0. Table 4.5 dB while the
standard deviation of the imperfect power control is increased from 0 to 2. Fig.7 also shows that the traditional uniform user distribution modeling matches well with the 2-D Gaussian model. σ = 2.7 shows the optimized average number of slots per sector for the 27 cells WCDMA network as the
Eb I0
is increased from 5 dB to 10 dB and the standard deviation of imperfect power
control is increased from 0 to 2.5 dB
9
SIR threshhold in (dB)
8
7
6
5 5 10 15 20 Average number of slots per sector 25 30
Figure 4. Because of IPC. the number of slots per sector decreases signiﬁcantly to 1.6: Average number of slot per sector for perfect and imperfect power control analysis with a Spreading Factor of 64.

59 4.70 3.70 4.38 3.70 4.22 5.45 4.05 4.00 4.82 2.32 3.75 3.21 5.20 115.5 7.60 3.23 3.04 4.5 5.49 4.60 4.59 4.82 2.45 3.69 104.82 2.44 3.20 5.59 5.93 87.02 4.60 4.84 3.24 Uniform User Distribution 0.84 3.60 5.31 3.83 3.93 3.05 4.10 5.25 4.78 2.09 5.36 3.5 5.06 3.5 7.37 3.36 2.93 3.72 3.35 2.69 4.77 4.59 3.70 3.5 dB.68 3.75 3.67 3.67 3.75 3.46 3.40 3.09 5.10 5.32 3.83 3.25 4.39 3.88 2.16 4.35 2.59 4.85 3.77 2.09 5.72 2.31 3.68 3.92 3.84 3.76 3.Table 4.04 4.72 3.46 3.10 5.45 3.88 2.35 3.89 2.40 3.92 3.69 4.36 3.70 4.02 3.45 3.84 3.38 3.50 4.82 3.85 3.88 2.22 5.38 3.44 4.10 5.59 4.0 7.21 116.74 3.85 3.25 3.84 2.82 3.46 3.69 4.31 3.77 2.02 4.60 4.25 4.77 2.59 3.32 3.78 2. σ: cef f Cell1 Cell2 Cell3 Cell4 Cell5 Cell6 Cell7 Cell8 Cell9 Cell10 Cell11 Cell12 Cell13 Cell14 Cell15 Cell16 Cell17 Cell18 Cell19 Cell20 Cell21 Cell22 Cell23 Cell24 Cell25 Cell26 Cell27 Network Capacity Average Capacity per Sector 2-D Gaussian function 0.59 4.31 3.48 4.35 3.23 4.85 3.93 3.20 5.68 3.78 2.70 3.78 2.77 3.89 2.72 3.09 5.67 3.46 4.84 2.36 2.00 4.77 2.29 3.90 2.02 3.39 3.59 4.48 4.42 3.0 7.60 4.29 1.23 4.60 4.83 2.23 4.5: Capacity calculation for uniform user distribution with SF = 16 and
User Modeling with Imperfect Power Control.32 3.90 2.23
Eb Io
= 7.70 2.06 3.67 3.92 3.
42
.30 1.87 2.20 5.60 4.83 3.00 3.44 3.92 87.39 3.77 3.38 3.83 3.44 4.68 3.00 3.70 104.

σ = 2.5 dB PPC (Uniform). scenario with
Eb I0
= 7. σ = 0 dB IPC (Gaussian).7 Conclusions An analytical model has been presented for optimizing capacity in multi-cell WCDMA networks. The results also show that the determined parameters of the 2-dimensional Gaussian model matches well with traditional
43
.
4.5 dB while the standard deviation of the imperfect power control is
increased from 0 to 2.8 shows the optimized average number of slots per sector for the 27 cells WCDMA network as the
Eb I0
is increased from 5 dB to 10 dB and the standard
deviation of imperfect power control is increased from 0 to 2. Fig. σ = 0 dB IPC (Uniform).5 dB IPC (Gaussian).5 dB IPC (Uniform).5 dB. σ = 1.7: Average number of slot per sector for perfect and imperfect power control analysis with a Spreading Factor of 16. σ = 1.5 dB
9
SIR threshhold in (dB)
8
7
6
5 0 2 4 6 Average number of slots per sector 8 10
Figure 4.5 dB due to the imperfect power control. 4. we can have many low rate voice users or fewer data users as the data rate increases. As expected.5 dB. The results matches the previous trends for higher spreading factors. σ = 2.SF = 16 10 PPC (Gaussian).5 to 1. Numerical results show that the SIR threshold for the received signals is decreased by 0.

26 1.05 1.16 1.38 2.30 1.54 1.52 1.27 1.48 32.64 1.18 1.45 1.67 1.5 2.04 1.18 1.28 1.77 1.09 1.45 1.55 1.76 1.48 1.45 1.23 1.07 1.07 1.02 1.30 1.76 1.44 1.43 1.64 1.26 1.21 1.43 1.45 1.33 1.76 1.76 1.09 1.80 1.02 1.27 1.71 1.80 1.44 1.07 1.23 1.80 40.05 1.24 1.27 1.15 1.09 1.02 1.45 1.64 1.04 1.04 1.02 1.23 1.16 1.76 1.02 1.45 1.Table 4.24 1.07 1.48 1.67 1.0 2.23 1.48 1.44 1.18 1.68 1.49 1.54 1.80 1.03 1.18 1.44 1.25 1.55 1.18 1.18 1.68 37.09 1.43 1.67 37.24 1.64 1.27 1.68 1.71 1.27 1.23 1.43 1.20 1.48 1.28 1.24 1.68 1.55 1.04 1.18 1.27 1.0 2.16 1.15 1.27 1.
44
.20 1.18 1.52 1.64 1.05 1.33 1.27 1.27 1.27 1.92 1.80 40.43 1.18 1.21 1.05 1.27 1.64 1.45 1.64 1.54 1.67 1.48 1.80 1.54 1.16 1.80 1.5 2.5 dB. σ: cef f Cell1 Cell2 Cell3 Cell4 Cell5 Cell6 Cell7 Cell8 Cell9 Cell10 Cell11 Cell12 Cell13 Cell14 Cell15 Cell16 Cell17 Cell18 Cell19 Cell20 Cell21 Cell22 Cell23 Cell24 Cell25 Cell26 Cell27 Network Capacity Average Capacity per Sector 2-D Gaussian function 0.15 1.20 1.48 1.54 1.22
Eb Io
= 7.43 1.45 1.32 1.26 1.22 Uniform User Distribution 0.20 1.02 1.21 1.25 1.38 2.15 1.26 1.30 1.54 1.6: Capacity calculation for uniform user distribution with SF = 4 and
User Modeling with Imperfect Power Control.76 1.54 1.54 1.48 33.43 1.27 1.80 1.30 1.27 1.76 1.29 1.06 1.30 1.23 1.25 1.43 1.25 1.48 1.32 1.02 1.07 1.5 2.64 1.5 2.04 1.33 1.27 1.55 1.02 1.

5 dB IPC (Uniform). σ = 0 dB IPC (Gaussian). σ = 0 dB IPC (Uniform).5 3 3. σ = 1. σ = 2.5 dB
9
SIR threshhold in (dB)
8
7
6
5 0. Our method of optimizing capacity is fast.5 2 2.5 dB PPC (Uniform).5 1 1. σ = 1. σ = 2. methods for modeling uniform user distribution. accurate.8: Average number of slot per sector for perfect and imperfect power control analysis with a Spreading Factor of 4.SF = 4 10 PPC (Gaussian).
45
.5 Average number of slots per sector 4 4.5
Figure 4. and can be implemented for large multi-cell WCDMA networks.5 dB IPC (Gaussian).

n1. nM. The existence of CAC algorithms in mobile phone systems protects the cellular network and users while achieving network performance objectives.. .g vg +
g=1 j=1.... which meet the E0b constraint..j=i g=1
nj.g vg κji. In this chapter. Rg τg Sg /N0
(5..
(5...3 that the number of calls in every cell must satisfy
G M G
ni. A CAC algorithm is considered global if the decision to admit a call is based on the current total number of calls in the entire network.1 .1 ..G call conﬁguration or a feasible state. satisfying the above equations is said to be in feasible nM. I 46
. we design a CAC algorithm for multi-cell WCDMA networks and analyze its complexity.
5. and performance.2 Feasible States Recall from section 4. decide whether a new connection could be admitted without impacting the quality of service (QoS) of current connections in a network. A global CAC is most often a centralized scheme while a local CAC is distributed in nature.2)
n1.G A set of calls n = . implementation..M. Each approach has advantages and disadvantages. i.1)
where cef f =
(g)
W 1 Rg − ∗ .1 Introduction Call admission control (CAC) algorithms. and local if the algorithm considers only a single cell for making that decision. call blocking rate. in general.e. CAC algorithms must be designed to fulﬁll a grade of service (GoS). In addition.CHAPTER 5 WCDMA CALL ADMISSION CONTROL AND THROUGHPUT 5.g − vg ≤ cef f
(g)
for i=1.. ..

g ’s and qij.... if qi. .Denote by Ω the set of feasible states... ni.g the probability that a call with service g in progress in cell i departs from the network..
5. n1. is in Bi.. Deﬁne the set of blocking states for service g in cell i as Bi... . .
(5. We denote by qi.g as the probability that a call with service g in progress in cell i remains in cell i after completing its dwell time. This mobility model is attractive because we can easily deﬁne diﬀerent mobility scenarios by varying the values of these probability parameters [65]. by varying qii..g = 0 (∀ g ∈ G).. .. then the average dwell time of a call of the same service in the network will be constant regardless of where the call originates and what the values of qii. .. and it is independent of earlier arrival times.g independent of other call arrival processes.g are.g is constant for all i and g.. then qij..3)
If a new connection or a handoﬀ connection with the service g arrives to cell i.. a new dwell time that is independent of the previous dwell time begins immediately.3 Mobility Model The call arrival process with service g to cell i is assumed to be a Poisson process with rate λi..1
.g be the probability that a call with service g in progress in cell i after completing its dwell time goes to cell j. In this case.G .g.
47
. throughput). nM...1 . Let qij. ni. ni. call durations and elapsed times of other users.G
∈Ω . attempt a handoﬀ to an adjacent cell.. If cells i and j are not adjacent.g ’s for a service g. it is blocked when the current state of the network. Thus. .g =
n∈Ω:
n1.. n. For example. At the end of a dwell time a call may stay in the same cell..g . . n1...g . .. or leave the network..... The call dwell time is a random variable with exponential distribution having mean 1/µ. Deﬁne qii.G .g and qij. nM.g ..1 .. we can obtain low and high mobility scenarios and compare the eﬀect of mobility on network attributes (e.g
... nM.

g ). Let Ai be the set of cells adjacent to cell i. µ.g .g be the handoﬀ rate out of cell j oﬀered to cell i for service g. Let νji. n). where ρj. and let p(ρ.g . (5. n) =
G (ρ /µ )nk..6) can be rewritten as νji.7)
νxj. .g + (1 − Bj.g = λj. µ. λj.g vg +
g=1 j=1. The new call blocking
p(ρ.g )qji.g µ(1 − qii.g k.g . is given by Bi.g − vg ≤ cef f .g = ν(Bj.
(5.We assume that the occupancy of the cells evolves according to a birth-death process.g nk. µ. νji. and the departure rate from cell i when the network is in state n is ni.g is the sum of the proportion of new calls of service g accepted in cell j that go to cell i and the proportion of handoﬀ calls with service g accepted from cells adjacent to cell j that go to cell i. n) = 1.g ) = (1 − Bj.6)
Equation (5.g vg κji. Bi.
n∈Ω
where P0 is a normalizing constant such that probability for service g in cell i.j=i g=1
nj.
(5.g = ρ(v. qji.g =
n∈Bi. µ the matrix of departure rates. Aj ) = λj.g )qji. is given by ρj. ρj. Thus
νji.g +
x∈Aj
(5.g k..g .g µi. n) be the stationary probability that the network is in state n.4)
(g)
for i = 1. where the total arrival rate or oﬀered traﬃc for service g to cell i is ρi.g ρj.g
x∈Aj
νxj.g )qji. M. The distribution p is obtained as follows
P0 0
p(ρ.g (1 − Bj. µ. otherwise.g . the total oﬀered traﬃc to cell j for service g.g .g .
(5.g . Let ρ be the matrix of oﬀered traﬃc of service g to the cells.g ..g
p(ρ.5)
This is also the blocking probability of handoﬀ calls due to the fact that handoﬀ calls and new calls are treated in the same way by the network.g = ni.g ! k=1 g=1
M
G
M
G
ni.8)
48
.

. the handoﬀ rates as a function of the blocking probabilities and the oﬀered traﬃc. we use an iterative method to solve the ﬁxed point equations. ..and where v denotes the matrix whose components are the handoﬀ rates νij for i. The total oﬀered traﬃc can be obtained from a ﬁxed point model [32].. i.. which describes the oﬀered traﬃc as a function of the handoﬀ rates and new call arrival rates. We then calculate the new values of the handoﬀ rates and repeat. 41. i. the number of calls in progress in all the cells of the network. In order to simplify the CAC algorithm. the number of calls in progress in the current cell.e. We calculate the oﬀered traﬃc by adding the given values of the arrival rates to the handoﬀ rates. the iterative approach converged to a unique solution. The questions of existence and uniqueness of the solution and whether the iterative approach in fact converges to the solution (if a unique solution exists) are generally diﬃcult to answer due to the complexity of the equations involved. Furthermore. 37. Since the cardinality of Ω is O(cef f M ∗G ). A call arriving to cell i with service g is accepted if and only if the new state is a feasible state. the probability of each state in the feasible region needs to be calculated.M. 67. For a given set of arrival rates. the calculation of the blocking probabilities has a computational complexity that is exponential in the number of cells combined with number of available services. We deﬁne an initial value for the handoﬀ rates. Clearly this CAC algorithm requires global state.e. and the blocking probabilities as a function of the oﬀered traﬃc. in all the numerical examples we solved. we consider only those CAC algorithms which utilize local state.4 WCDMA Call Admission Control A CAC algorithm can be constructed as follows. 68].
5. 66. to compute the blocking probabilities. The blocking probabilities are now calculated using the oﬀered traﬃc. To this end we
49
. j = 1. Kelly has shown that for ﬁxed alternate routing the solution to the ﬁxed point problem is in fact not unique [36]. This approach has been extensively utilized in the literature to obtain solutions of ﬁxed point problems [4.

g − λi.11)
where B is the vector of blocking probabilities and λ is the matrix of call arrival rates. We note that the complexity to calculate the blocking probabilities in (5. . Once the maximum number of calls with diﬀerent service that are allowed to be admitted in each cell.
(5.10)
Ak /k! i.g = B(Ai. and the bit error rate requirement is guaranteed since Ω ⊂ Ω. Clearly the set of admissible states denoted Ω is a subset of the set of feasible states Ω. To study the eﬀect of mobility and to diﬀerentiate between new calls and handoﬀ calls.g /µi. T .g k=0
..g Ai.g = ρi. N. the throughput function can be generalized to a revenue function.deﬁne a state n to be admissible if ni.. .g is a parameter which denotes the maximum number of calls with service g allowed to be admitted in cell i.g /µ(1 − qii.
(5..g (ρi. Ni. λ) =
i=1 g=1
{λi.5 Network Throughput The throughput of cell i consists of two components: the new calls that are accepted in cell i minus the forced termination due to handoﬀ failure of the handoﬀ calls into cell i for all services g. (5. the CAC algorithm for cell i for service g will simply compare the number of calls with service g currently active in cell i to Ni.g (1 − Bi.g ) − Bi.g /Ni. is calculated (this is done oﬄine and described in the next section).9)
where Ni. The blocking probability for cell i with service g is then given by
i.g ≤ Ni. G. ρ.. Hence the total throughput. M and g = 1.g
where Ai = ρi.
5.g !
N
Bi.. Thus our CAC algorithm is implemented with a computational complexity that is O(1).g .g in order to accept or reject a new arriving call.g ) is the Erlang traﬃc in cell i with service g.g )} . The term revenue suggesting 50
.g for i = 1. of the network is
M G
T (B.g ) =
Ni..10) is O(M ∗ G).

Hence the revenue.
(5.6 Calculation of N We formulate a constrained optimization problem in order to maximize the revenue subject to upper bounds on the blocking probabilities and a lower bound on the signal-to-interference constraints in (5.
5. H. for all the users for diﬀerent services g. In this optimization problem the arrival rates are given and the maximum number of calls that can be admitted in all the cells are the independent variables.g ) − ci. This is given in the following
51
.g control the tradeoﬀ between new calls and handoﬀ calls.g is the cost of a forced termination of a call with service g due to a handoﬀ failure in cell i. Surveys and market studies investigating these issues help network administrators set these weights and achieve a desired tradeoﬀ.g (ρi. For example assume that ri. becomes
M G
H(B. These are tools that allow the system administrator to place more importance on handoﬀ calls.g (1 − Bi.g and ci.g λi.g = cg for all i. Then.g = 1 and ci. it is easy to see that for a given network topology the choice of larger cg for service g will tend to increase the values of N for those cells into which the handoﬀ rates are high while decreasing the values of Ni. βg .g is the revenue generated by accepting a new call with service g in cell i. ρ. The goal is to optimize the utilization of network resources and provide consistent GoS while at the same time maintaining the QoS.g Bi.1).g )} .g for the other cells.g − λi. λ) =
i=1 g=1
{ri.an economic meaning is chosen to emphasize the rewards from not blocking a new call and the penalty (whether measured monetarily or by customer aggravation) from having handoﬀ calls blocked.g aﬀect the CAC algorithm through their eﬀect on N. The values of ri. and ci.12)
where ri.g and ci. The choice of ri.

g ) ≤ βg . λ). The optimization problem in (5.g vg +
g=1 j=1.g vg +
g=1 j=1.. . The base stations are located at the centers of a hexagonal cell whose radius is 1732 m. ρ..G
H(B. This is given in the following max
λ..
G M G
subject to
Ni.
G M G
Ni. 52
.13)
(g)
for i = 1.g − vg ≤ cef f .14) provides an upper bound on the total throughput that the network can carry...13) is solved oﬄine to obtain the values of N.
subject to
B(Ai. NM... B(Ai.. Base station 1 is located at the center of the network. NM. M. (5. the base stations are numbered consecutively in a spiral pattern.g vg κji.
The optimized objective function of (5.. .G N= .. N1.g . . (5.1 .1 .
5..max N1.g ) ≤ βg .. ρ. Ni.g ...
5. λ)..14)
(g)
for
i = 1.7 Maximization of Throughput A second optimization problem can be formulated in which the arrival rates and the maximum number of calls that can be admitted in all the cells are the independent variables and the objective function is the throughput. M.8 Numerical Results The following results have been obtained for a 27-cell CDMA network. This is the network capacity for the given GoS and QoS.j=i g=1
Nj. Ni.g vg κji..j=i g=1
Nj.g − vg ≤ cef f . N
T (B. .

and high mobility of users are considered. In all three mobility scenarios. 16.240 0.010 qii. In the numerical results below. The interference-to-background-noise ratio is 10 dB.375. 64. Thus.700 0. and 256. and average mobile height of 1.012 0.g = 0. respectively.
Ai 3 4 5 6 qij. We assume that the mobility characteristics for a given service g stays the same throughout diﬀerent cells in the network.700
The COST-231 propagation model [49] with a carrier frequency of 1800 MHz. we analyze the average throughput per cell by dividing the results from (5.g 0. The activity factor is 0.700 0. Tables 5.06 dB.5 dB with perfect power control for all four scenarios.240 0.g 0.1: The low mobility characteristics and parameters.14) by the total number of cells in the network and multiplying by the maximum data rate in Table 4.02 dB.020 0.04 dB. average base station height of 30 m.240 0.2 show respectively the mobility characteristics and parameters for the low and high mobility cases.
53
.08 dB for Spreading Factor of 4.240 qi.7. and 24. 12. The bit-energy-to-interference ratio threshold is 7. the average dwell time of a call in the network is constant. The following parameters are used for the no mobility case: qij. for each SF value. qii. The path-loss coeﬃcient is 4. The shadow fading standard deviation is 6 dB and the processing gain is 6.700 0.g = 0. regardless of where the call originates and mobility scenario used.Table 5.g = 0.g 0.7 for all cells i and j.3 and qi. Three mobility scenarios: no mobility. low mobility.5 m is used to determine the coverage region.015 0. the probability that a call leaves the network after completing its dwell time is 0. 18.2.1 and 5.

which is used for voice communication as shown in Fig. The results for the average throughput for no mobility and high mobility cases are almost identical while the throughput for low mobility is higher for each blocking probability.g 0. 5. 4.
• qii.3 shows the optimized values of N for each cell for all three mobility models with perfect power control and 2% blocking probability. The low mobility case has an equalizing eﬀect on traﬃc resulting in slightly higher throughput.g 0 0 0 0 qi. which are adjacent to cell i.
5. the number of possible concurrent connections within one cell is also decreased.050 qii.2: The high mobility characteristics and parameters.1 shows the optimized throughput per cell for a blocking probability from 1% to 10%. which is used to carry data for the control channel.g 0. Fig.1 0.
Ai 3 4 5 6 qij.8. Table 5. we set SF equal to 256.2 WCDMA Throughput Optimization with SF of 64 Next.3. we set SF equal to 64. the lower trunking eﬃciency [48] leads to
54
.700
•
Ai
is the number of cells. Because the throughput is calculated based on the number of simultaneous connections between MSs and BSs.1 WCDMA Throughput Optimization with SF of 256 First.075 0.8.g is the probability that a call with service g in progress in cell i remains in cell i after completing its dwell time.g is the probability that a call with service g in progress in cell i after completing its dwell time goes to cell j.g is the probability that a call with service g in progress in cell i departs from the network.060 0.700 0.Table 5.
5.700 0. As a result of lowering the SF to 64.700 0.
• qi.
• qij.

29 62.71 50.83 71.02.43 73.82 73.84 71.85 53.73 50.27 62.43 71.85 51.85 53.73 62.84 51.
No Mobility Cell ID Cell1 Cell2 Cell3 Cell4 Cell5 Cell6 Cell7 Cell8 Cell9 Cell10 Cell11 Cell12 Cell13 Cell14 Cell15 Cell16 Cell17 Cell18 Cell19 Cell20 Cell21 Cell22 Cell23 Cell24 Cell25 Cell26 Cell27 Ni 52.40 Low Mobility Ni 52.73 53.95 51.84 51.84 53.84 53.73 50.84 51.71 50.84 71.86 73.84 53.Table 5.01 50.40 71.73 50.40 71.71 63.00 50.00 50.95 51.95 51.43 73.73 62.83 71.43 71.73 62.27 62.00 50.73 62.74 73.43 71.74 63.74 63.82 73.71 50.71 63.43 73.95 51.95 51.86 53.86 53.74 73.85 51.86 73.40
55
.74 73.29 62.82 73.73 53.73 53.3: Calculation of N for uniform user distribution with SF = 256 and blocking probability = 0.85 51.85 53.01 50.73 62.74 63.01 50.40 High Mobility Ni 52.83 71.40 71.27 62.95 51.84 71.29 62.73 62.71 63.86 73.86 53.

8.06 0.30 as was shown in Table 4.8.SF=256
Average Throughput per Cell (kbps per unit time)
290 285 280 275 270 265 260 255 250 0.1: Average throughput in each cell for SF = 256.1
Figure 5. The resulting throughput. the average number of slots within one cell decreases to 4. As a result of lowering the SF to 16.03 0. lower throughput as shown in Fig.4 WCDMA Throughput Optimization with SF of 4 Next. we set SF equal to 4. we set SF equal to 16.08 0. Table 5.01 0. Table
.8. the average slots per sector decreases signiﬁcantly to 1.05 0.
5.02 0.4 shows the optimized values of N for each cell for all three mobility cases and SF equal to 64. As a result of lowering the SF to 4. is much lower compared to the case with SF equal to 64 or 256.3 WCDMA Throughput Optimization with SF of 16 Next.5 dB as shown in Fig. which is used for 64 Kbps data communication as shown in Fig.5. Table 5.
5. as shown in Fig.07 Blocking probability threshold CAC high mobility CAC low mobility CAC no mobility 0. 5.2. 4. 4.04 0.09 0.49 with perfect power control and 56
Eb I0
= 7.3. 5.5 shows the optimized values of N for each cell for all three mobility cases with SF equal to 16. which is normally used for 256 Kbps data communication between BSs and MSs.4.

SF=16
Average Throughput per Cell (kbps per unit time)
170 160 150 140 130 120 110 100 90 0.1
Figure 5.06 0.09 0.03 0.06 0.04 0.05 0.1
Figure 5.08 0.2: Average throughput in each cell for SF = 64.SF=64 250
Average Throughput per Cell (kbps per unit time))
240
230
220
210
200
190
180 0.02 0.05 0.04 0.01 0.01 0.3: Average throughput in each cell for SF = 16.08 0.
57
.02 0.09 0.07 Blocking probability threshold
CAC high mobility CAC low mobility CAC no mobility 0.07 Blocking probability threshold CAC high mobility CAC low mobility CAC no mobility 0.03 0.

86 18.03 13.25 16.45 18.32 13.32 13.11 13.86 13.11 16.58 13.61 13.11 13.11 16.11 16.86 18.32 13.03 16.25 16.86 13.11 16.85 18.32 13.26 16.58 13.62 13.
No Mobility Cell ID Cell1 Cell2 Cell3 Cell4 Cell5 Cell6 Cell7 Cell8 Cell9 Cell10 Cell11 Cell12 Cell13 Cell14 Cell15 Cell16 Cell17 Cell18 Cell19 Cell20 Cell21 Cell22 Cell23 Cell24 Cell25 Cell26 Cell27 Ni 13.45 18.03 16.85 18.62 13.86 13.45 18.62 13.85 Low Mobility Ni 13.11 13.4: Calculation of N for uniform user distribution with SF = 64 and blocking probability = 0.45 18.86 18.26 16.86 13.03 16.86 18.32 13.46 18.11 16.86 13.45 18.32 13.45 18.26 16.85 18.86 18.32 13.03 16.45 18.58 13.03 18.61 13.03 13.11 16.03 18.85 High Mobility Ni 13.03 13.46 18.86 18.25 16.02.03 16.32 13.86 13.11 13.11 13.03 18.03 16.85
58
.32 13.32 13.61 13.32 13.32 13.45 18.11 13.45 18.46 18.Table 5.

22 5.60 4.60 3.10 5.60 5.60 5.10 5.Table 5.46 3.21 5.50 4.83 3.68 3.46 4.83 3.68 3.22 5.49 4.68 3.10 5.60 4.83 3.68 3.46 3.10 5.02.76 3.77 3.76 3.21 5.10 5.83 3.68 3.45 3.60 4.45 4.77 3.68 3.10 5.75 3.10 5.22 5.68 3.10 5.49 4.68 3.68 3.46 3.60 4.46 4.60 5.21 5.60 4.77 3.22 5.45 4.45 3.
No Mobility Cell ID Cell1 Cell2 Cell3 Cell4 Cell5 Cell6 Cell7 Cell8 Cell9 Cell10 Cell11 Cell12 Cell13 Cell14 Cell15 Cell16 Cell17 Cell18 Cell19 Cell20 Cell21 Cell22 Cell23 Cell24 Cell25 Cell26 Cell27 Ni 3.50 4.46 4.75 3.75 3.10 5.68 3.45 3.21
59
.60 4.68 3.22 5.68 3.10 5.21 Low Mobility Ni 3.45 4.83 3.76 3.5: Calculation of N for uniform user distribution with SF = 16 and blocking probability = 0.22 5.10 5.21 High Mobility Ni 3.83 3.49 4.50 4.60 3.10 5.60 3.

every connection moves to a new cell.09 0.1
Figure 5. Numerical results show that as the spreading factor increases.03 0. Thus.SF=4
Average Throughput per Cell (kbps per unit time)
90 80 70 60 50 40 30 20 0.4: Average throughput in each cell for SF = 4. the optimized throughput is better.01 0.08 0. due to the trunking eﬃciency for all the three mobility models.05 0.4. For the high mobility case.5 shows the optimized values of N for each cell for all three mobility models with SF equal to 4.
5. However. which is equivalent to having new calls in every cell for each dwell time.06 0.9 Conclusions An analytical model has been presented for CAC algorithm for optimizing the throughput in multi-cell WCDMA networks. 60
. The average throughput for all three mobility cases are almost identical as shown in Fig.02 0. Our CAC algorithm is computationally eﬃcient and can be implemented for large multi-cell WCDMA networks. the throughput for the low mobility scenario is slightly higher because of the equalizing eﬀect on traﬃc due to low mobility.04 0. 5. 5. the throughput for the no mobility and high mobility scenarios are identical.07 Blocking probability threshold
CAC high mobility CAC low mobility CAC no mobility 0.

82
61
.27 1.76 1.02.32 1.
No Mobility Cell ID Cell1 Cell2 Cell3 Cell4 Cell5 Cell6 Cell7 Cell8 Cell9 Cell10 Cell11 Cell12 Cell13 Cell14 Cell15 Cell16 Cell17 Cell18 Cell19 Cell20 Cell21 Cell22 Cell23 Cell24 Cell25 Cell26 Cell27 Ni 1.83 1.93 1.80 1.92 Low Mobility Ni 1.27 1.54 1.79 1.58 1.80 1.59 1.82 1.92 1.24 1.94 0.55 1.54 1.48 1.42 1.79 1.94 1.27 1.Table 5.59 0.30 1.6: Calculation of N for uniform user distribution with SF = 4 and blocking probability = 0.76 1.27 1.37 0.58 0.53 1.83 1.24 1.53 1.55 1.48 1.28 1.79 1.58 1.93 0.42 1.53 1.93 1.32 1.82 1.30 1.58 0.28 1.76 1.53 1.30 0.16 1.94 0.79 1.30 1.54 1.93 1.32 1.92 1.93 1.80 1.23 1.92 1.94 1.80 1.82 1.23 1.93 1.23 1.32 1.76 1.32 1.83 1.54 1.83 High Mobility Ni 1.28 1.59 1.37 0.59 0.28 1.

This chapter summarized the features. In Chapter 2. This allowed for the calculation of the intra-cell and inter-cell interference. CDMA is an unique multiple access technology. The advantage gained from this model over actual interference is a huge reduction in computational complexity. which divide the total bandwidth to frequency sub-bands and time slots.1 Summary Compared to other multiple access technologies like FDMA and TDMA. In Chapter 3. Interference generated by users in the network is the main limitation for determining the network capacity. we provided an overview of CDMA and WCDMA. which demonstrated CDMA’s superiority over other 2G cellular networks. In addition. this chapter also introduced the requirements for 3G cellular systems. which can be used to optimize the placement of BSs and radio network controllers as well as to analyze the performance of resource management algorithms towards meeting the ﬁnal goal: the calculation and maximization of network capacity. as well as new additional features. We analyzed three scenarios of
62
. where all users share the entire available bandwidth to transmit and receive information. respectively. which included wider bandwidth allocations. User modeling helps compute the traﬃc density in the cellular network. we presented an analytical model for approximating the user distributions in multi-cell WCDMA networks using 2dimensional Gaussian function by determining the means and the standard deviations of the distributions for every cell.CHAPTER 6
CONCLUSIONS 6. and the reverse-link capacity of the network. A large number of researchers in CDMA use average interference for designing call admission control and calculating capacity.

6. To achieve the promised speed of 2 Mbps data rate in 3G systems.
63
. The results showed that as the spreading factor increases.2 Future Research We conclude by outlining possible directions for future research: As shown in Table 4. and 4. we designed a call admission control algorithm for optimizing the throughput in multi-cell WCDMA networks. low. In Chapter 5. Numerical results showed that the SIR threshold for the received signals was decreased by 0. In Chapter 4. and high mobility scenarios. 6 parallel codes with a spreading factor of 4 need to be used. 64.2. and how the interference level is generated and capacity is calculated in each cell in WCDMA networks with 2x60 MHz spectrum (1710-1770 MHz and 2110-2170 MHz). The new IMT-2000 spectrum of 190 MHz (2500-2690 MHz) arrangement is still under development. 16.user distribution using 2-D Gaussian function: uniform distribution of users. and users distributed at cells’ boundaries. we can have many low rate voice users or fewer data users as the data rate increases. the maximum transfer data rate is only 480 Kbps when used with a spreading factor of 4 in 5 MHz spectrum. Optimizing capacity was calculated with diﬀerent spreading factors of 256. As expected. Numerical results were analyzed for no.5 dB due to the imperfect power control. the optimized throughput is better for all the three mobility model. users densely clustered at the center of the cells.5 to 1. we maximized the capacity in multi-cell WCDMA networks. The results also showed that the determined parameters of the 2-dimensional Gaussian model matched well with the traditional method for modeling user distribution. Further research can delve into how to use and interact between these parallel codes. due to trunking eﬃciency. Our method of optimizing capacity and throughput is computationally eﬃcient and can be implemented for large multi-cell WCDMA networks.

64
. Akl. Akl.. CCAP: CDMA Capacity Allocation and Planning.G.G. 50(3):711–722. Akl. Akl. and K. IEEE Trans. May 2001. Signal and interference statistics of a CDMA system with feedback power control.S. Technol. and M.
http://www. Technologies and Applications.. N. 54:639–651. Ariyavisitakul and L. [7] S. [2] R. and P. M. IEEE Wireless Commun. [4] R. and P. Min.V. Global Telecommunications Conf.htm. Chockalingam. and Networking Conf. IEEE Trans. Hegde. Outage and capacity analysis of cellular CDMA with admission control.G. Naraghi-Pour. Veh. Hegde. July 2004. 3GPP speciﬁcations releases. M. Technol. Anand. [3] R. Chang.. 3:404–408. Min. and P. Akl and A. M. Washington University.. Parvez. June 1999. Eﬀects of call arrival rate and mobility on network throughput in multi-cell CDMA. and stages. IEEE International Conf. Mobility-based CAC algorithm for arbitrary traﬃc distribution in CDMA cellular systems.G. phases. March 2005. M. M. Technical report.V. A.F. A. Min. March 2002. [5] R. Proceedings of SCI 04: Communication and Network Systems.. Veh. Hegde. April 1998. Naraghi-Pour.BIBLIOGRAPHY [1] 3GPP.G. [6] R. Impact of interference model on capacity in CDMA cellular networks. Hegde. December 1991.S. Sivarajan. 2:1490–1495. Multi-cell CDMA network design. Chandra. 3:1763–1767. 2:908–912. [8] S.3gpp.org/specs/releases. on Commun.V.S.

IEEE Veh.asp?h area=0.org/worldwide/index. CDG : Worldwide : CDMA Worldwide. of the IEEE Computer and Commun. [10] A. Priscoli. July-August 1996. [11] P. Salz. [17] M.cdg. Asawa and W. European Trans. Capacity gain of an uplink-
synchronous WCDMA system under channelization code constraints. [16] CDMA Development Group. and F. and J. Analysis of CDMA range extension due to soft handoﬀ. Salz. Rohani.com/library/weekly/aa070899. May 1992. Technol. Carnero. Balaban and J. Commun. July 2004. http://www.I.htm. 7(4):305–314. Theoretical considerations.[9] M.
65
. Optimal scheduling of soft handoﬀs in DS/CDMA communication systems. IEEE Trans. April 1995. Optimum diversity combining and equalization in digital data transmission with applications to cellular mobile radio I.about.asp. July 1995. Societies.. Eﬀects of user mobility on the capacity of a CDMA network. K.E. Sestini. Stark. CDG : Worldwide : 4Q 2004 CDMA Subscribers Statistics. 53:982 – 991. Selling the cell phones history of cellular phones.cdg.org/worldwide/cdma world subscriber..D. Mogensen. [13] Mary Bellis. Fourteenth Annual Joint Conf. Conf. [15] CDMA Development Group. Pedersen. 2:917 – 921. [12] P.
http://inventors.O. IEEE Veh.E. Conf. Optimum diversity combining and equalization in digital data transmission with applications to cellular mobile radio II. Chopra. IEEE Trans. 40:895–907. Balaban and J. 40:885–894. Baiocchi. [14] J. 1:105 – 112. Reed. Commun. and P. Technol.. on Telecommunications.
http://www. Numerical results. F.. K. May 1992.

3G Mobile Commun. Aghvami. L.A. January 1997. IEEE Trans. Lett. A simple distributed autonomous power control algorithm and its convergence.. Sawahashi. Khanifar.J.. A.. Prentice Hall communications engineering and emerging technologies series. [23] G. Fu and J. and F. January 1999.S. Garg. Weaver. [25] V.. IEEE International Conf. and C. 42(4):641–646. a mobility model with a provider-oriented approach. Fitkov-Norris and A. and A.[18] G. Technol. November 1995. 40(2):303– 312. Adachi. 5:2835–2839. [26] K.H. on Universal Personal Commun. Electr.. Veh. [20] J.S. Technol. Tree-structured generation of orthogonal spreading codes with diﬀerent lengths for forward link of DS-CDMA mobile radio. Performance of SIR based power control in the presence of non-uniform traﬃc distribution. Technol.D. and Okawa K. [21] Adachi F. Dohi.. Everitt.E. pages 63 – 67. 48(1):153–165. Performance evaluation of a two layer hierarchical cellular system with variable mobility users using multiple class applications. Technol.M. Evans and D. May 2002. IS-95 CDMA and CDMA2000: Cellular/PCS Systems Implementation. IEEE Veh. Veh. Dynamic pricing in cellular networks. Technol. M. [19] T. Viterbi. I. pages 534–538.. Padovani. pages 334–338. IEEE Trans. [22] E. On the teletraﬃc capacity of CDMA cellular networks. November 1993. Miljanic. September 1999. [24] H. 33:27–28.. Sawahashi M. Conf. Jacobs. Technol. Weerakoon.. March 2001.J.D. Wheatley. R. D. 3G Mobile Commun. Veh. Cimone.S. May 1991. Downlink capacity analysis in 3GPP WCDMA networks system. Thompson. On the capacity of a cellular CDMA system. IEEE Trans..
66
. Gilhousen. Foschini and Z. 2000.K.

1981. 44(2):247–256.[27] S. Hanly.asp?doc id=15195&page number=9.A. [36] F. Reidel. Areas Commun.H. Hanly. 1986.M. [35] V. Local Computer Networks. Technol.unstrung. Cuberovic. Advances in Applied Probability. 15(8):1627–1635. Air interfaces what are they?
http://www. Capacity and power control in spread spectrum macrodiversity radio networks. Conf. Copeland. Holma and A.. Jovanovic and D. 3:1757 – 1762.html. An algorithm for combined cell-site selection and power control to maximize cellular spread spectrum capacity.
http://www. Commun.infosyncworld. Areas Commun. John Wiley & Sons. Umeda. 2002. [29] H. infoSync World : 1.. Oct 1998. October 1997. Blocking probabilities in large circuit switched networks.. IEEE Trans. Kelly. Istratescu. Theoretical and experimental analysis of the new soft handoﬀ algorithm for CDMA systems.
67
. [32] V. [33] S.I. Select..com/news/n/5048. Jeong and J. February 1996. Toskala. Fixed Point Theory : An Introduction. [28] S.V. Cell loss ratio and multiplexing gain of an ATM multiplexer for VBR voice sources. WCDMA for UMTS: Radio Access for Third Generation Mobile Communication. [34] Dan Jones.87 billion mobile users by 2007. Select. IEEE J. D. Capacity design and performance of call admission control in cellular CDMA systems..com/document. IEEE J. September 1995. IEEE Veh. May 2000. 18:473–505. Ltd.P. 13(7):1332–1340. pages 384–389. [31] Y. Ishikawa and N. [30] infoSync World.V.

1988. IEEE/ACM Trans. J. Proceedings of the 35th Southeastern Symposium. Liu. Technol. Eﬀect of soft handoﬀ parameters and traﬃc loads on soft handoﬀ ratio in CDMA systems.P. [38] Kyoung II Kim. July 2005. Advances in Applied Probability. November 1993. Kelly. April 2003. Castro. Morrison. Systems and Computers. Capacity improvement of base-station antenna arrays cellular CDMA. and K. Ojanpera. 2:782 – 785. 4(4):531–543. International Conf. J. ATM network design and optimization: a multirate loss network framework. pages 418–152. Kordybach.A. 2:490–494. [39] Z. Wang. Girard. on Signals. K. on Comput.[37] F. on Networking. Kemppi. User distribution and mobility model framework for cellular network simulations. and Control Technol. pages 518–522. Yang. Matolak and Anjula Thakur.F. [45] T. IST Mobile & Wireless Telecommunications Summit. Kailath.. August 1993.. A. 2:1437–1441. Conf. Klein.
68
. Commun. Record of The Twenty-Seventh Asilomar Conf. May 1997. [41] D. Naguib. on Communication Technology Proceedings. and T. June 2002. 42:345– 350. Nousiainen. and A. accepted. Paulraj. J. Technol.. 20:112–144. L. and D. and P. Y. Approximating user distributions in WCDMA networks using 2-D Gaussian. Comparison of multiple access schemes for UMTS. [42] A.. Skold.G Ramakrishnan. Mitra. IEEE Veh. IEEE Veh. Outside cell interference dynamics in cellular CDMA. [40] David W. [44] S. Akl. CDMA cellular engineering issues. August 1996. March 2003.. International Conf. Routing in circuit-switched network: Optimization. shadow prices and decentralization. Nguyen and R. Conf. Conf. [43] S.

Wireless Communications: Principles and Practice. using A new IS-95 class CDMA of precision base UTC and fre-
reference
station
transmissions. 2001. 1(3):28–34... 48(1):233–239. Ravi. Technol. Rappaport.U. on Personal Wireless Commun.K. Reverse link performance of IS-95 based cellular systems. IEEE International Conf.S.
69
. Rappaport. Third Quarter 1994. CDMA mobile systems with tailor made power control to each mobile station. IEEE Trans.com/pdf/PTTI2001 WhitePaper.
http://www. July 1998. Cho. Mag. [52] S. Siu and K. SIR-Based transmit power control of reverse link for coherent DS-CDMA mobile radio.. E81-B(7):1508–1516.[46] R. Adachi. and F. Veh. Penrod. August 1994. [54] Y. and D.V. C. Capacity enhancement for CDMA systems through adaptive cell sectorization. IEEE Personal Commun.M. Prentice Hall. IEEE Wireless Commun.. 3G Mobile Commun. Seo. Soo. Comparison of multiple-accessing schemes for mobile communication systems. [51] C. September 1999.. [53] S. Shin. January 1999. Dohi. [48] Theodore S. and Networking Conf. Padovani. Technol. March 2000. Yener. on Commun. Interference-based channel assignment for DS-CDMA cellular systems. pages 46–50. [47] Bruce quency M.pdf. [49] T. [50] K. Prentice Hall. T. Wireless Communications Principles and Practice.endruntechnologies. Saraydar and A.. pages 152–156. Sung. IEICE Trans. pages 1139– 1143. 1996.

. Approximating the othercell interference distribution in inhomogenous UMTS networks.J. [59] K. Leibnitz. Technol. Tachikawa. IEEE International Symposium on Spread Spectrum Techniques and Applications. A. February 1990. and Mobile Radio Commun. Tran-Gia. Paulraj. [60] K. Ltd. J. IEEE International Conf.F. Suard. Evaluation of a CDMA cell design algorithm considering nonuniformity of traﬃc and base station locations. Takeuchi. July 1998. Takeo and S. Technol.. IEEE International Conf. 1:229–233. Heck. 2002. Edwards. W-CDMA: Mobile Communications System. E81A(7):1367–1377. Weller. IEEE Veh. Diaz-Estrella. B.M. Results of multiple cell simulations of a CDMA cellular system focusing on SIR based power control.. 39:56 – 67. K. IEEE Trans.P. on Commun. Beach. Xu. May 2002. and J. April 1993. Romero-Jerez. [62] C. and P. IEICE Trans. [56] B. Conf. The proposal of CDMA cell design scheme considering change in traﬃc distributions. 4:1640–1644.C. [57] S. 1:63–67. Sato. 4:153 – 166. September 2002. G.A. Performance of CDMA mobile communication systems using antenna arrays. Capacity analysis of SIRbased power-controlled CDMA systems under multipath fading. and A. McGeehan. IEEE International Symposium on Personal. Fundamentals. The performance enhancement of multibeam adaptive base-station antennas for cellular land mobile radio systems. Indoor. and A. Takeo and S. Tellez-Labao. 3:1237–1241.[55] D.
70
. [61] Y. John Wiley & Sons. D. Swales. Naguib. Schroder. June 1998.. Speech. [58] K. on Acoustics. September 1998. A. Veh. K. and Signal Processing.. M. Sato. Staehle.

Technol. A. 12(8):1281–1288. M. and E.[63] TIA/EIA.J.S. Viterbi and A. IEEE J. March 1999. and M. Soft handoﬀ extends CDMA cell coverage and increases reverse link capacity.J. Naraghi-Pour. May 1995. Zehavi. Communication Network Design and Evaluation Using Shadow
Prices. IEEE J. M. Technical report. on Commun. A. PhD thesis.Wireless Networks. [64] TIA/EIA. Viterbi. Implied costs for multi-rate wireless networks. Zehavi. 3:1612–1616. Gilhousen. Erlang capacity of a power controlled CDMA system. Hegde. Select. K. Technical report.S. Blocking eﬀects of mobility and reservations in wireless networks. IS-95B.M.. Viterbi. IEEE J. Vargas. and E. Areas Commun. October 1994. [65] C.. [69] A. 2:904–908. Mobile station-base station compatibility standard for dual-mode wideband spread spectrum cellular system. Gilhousen. [66] C. 11(6):892–900. and M. August 1993. 10:323 337. 71
.J. CDMA Principles of Spread Spectrum Communication. 12:1281–1288.. Addison-Wesley. IS-95A. June 1998. Soft handoﬀ extends CDMA cell coverage and increases reverse link capacity..V. Areas Commun. Viterbi. 1996. 1995. Select.. May 2004. Naraghi-Pour. May 1998. [71] A.M. October 1994. Mobile station-base station compatibility standard for wideband spread spectrum cellular systems. Vargas. Hegde. Conf. Areas Commun. [68] C. J. Naraghi-Pour. K. IEEE Veh. [70] A. Vargas-Rosales. Vargas. Viterbi. IEEE International Conf. M. [67] C. and M. Select. Implied costs in wireless networks. Hegde.V. Louisiana State University.Viterbi.J.M. Viterbi. [72] A.

.. Conf. September 1995. Conf. IEEE Veh. A framework for uplink power control in cellular radio systems. Areas Commun. Technol. Worley and F. [76] Q.[73] J. Sept. 48:571–575.L. Wen. Select. J. Yates. J. Handoﬀ schemes in CDMA cellular systems. March 1999. UMTS air interface voice/data capacity-part 2: forward link analysis. Wong and T. and Signal Processing 1998.. May 2001. Proceedings of Commun. UMTS air interface voice/data capacity-part 1: reverse link analysis. December 1997. on Commun. 6:1738–1742.H. Chen. IEEE Veh. [78] Q. [77] R. Sheu. Mag. May 2001. Soft handoﬀs in CDMA mobile systems. Yue. pages 255 – 260.S. Performance of optimum transmitter power control in cellular radio systems. IEEE Personal Commun.
72
. Zhang. Veh. Lim. 1998. [74] D. June 2001.. IEEE J. Technol. 4:2730 – 2734. Wu. 4:2725 – 2729..D. 13(7):1341–1347. IEEE Trans. Zhang and O. [79] Q. [75] B. and J. IEEE International Conf. Technol.. 4:6–17. An optimum downlink power control method for CDMA cellular mobile systems. Takawira.