Dissertation Organization

In this introductory chapter we provide the synopsis of the thesis. This chapter presents the research motivation, paper survey, and thesis outline.

In Chapter 2, the evolution of the mobile systems is briefly reviewed. The importance of power control in the reverse link of a CDMA system is highlighted. Various kinds of the power control schemes in CDMA systems are described in this chapter. Among these power control schemes, the most important two schemes are the strength-based and the SIR-based power control schemes. The analysis of these two kinds of power control schemes is the scope of this dissertation.

In Chapter 3 we focus on the strength-based power control scheme. The system performance of a truncated closed-loop power control (TCPC) scheme for uplinks in DS/CDMA cellular systems over frequency-selective fading channels is conducted. It is shown that the TCPC scheme can achieve higher system capacity than the conventional strength-based power control scheme.

We also successfully analyze an imperfect TCPC scheme (or called realistic TCPC scheme), i.e.

TCPC scheme under realistic consideration of power control error due to power control step size, power control period, power control delay, and MS velocity. A closed-form formula is obtained to accurately approximate the system capacity of the imperfect TCPC scheme.

In Chapter 4 we turn to focus on one important factor, the interference statistics, in a strength-based power controlled DS/CDMA cellular system. The interference statistics of a DS/CDMA system are essential to the understanding of the system’s dynamics. Although the conventional SGA (standard Gaussian approximation) method is easy to use and applicable to a complicated circumstance, e.g. the cellular system over the frequency-selective fading channel in

[17], it is known that the SGA is not very accurate [18]. In order to improve accuracy while still applicable to a complicated circumstance, we propose an approximation method by characteristic function (AM-CF) to approximate the distribution of the MAI (multiple access interference) signals in DS/CDMA cellular systems. The method considers the effect of a frequency-selective multipath fading channel; it also assumes perfect power control and a rectangular/sinc chip waveform. Using this method, the distribution of the MAI signals is more accurately approximated.

In Chapter 5 we focus on another important kind of power control scheme ― the SIR-based power control scheme. In this chapter, we propose a novel analytical method to analyze the system capacity of an imperfect SIR-based power control scheme for uplinks in DS/CDMA cellular systems wherein the received SIR is a random variable. The system behavior can be described by a set of linear equations. A closed-form solution for the system capacity is successfully derived. And the derived formula needs only the first and second order statistics of each element in the coefficient matrix of the linear equation. Results show that the analytical and the simulation results are substantially matched together, which implies the novel analytical method is quite accurate.

Finally, concluding remarks and future research topics are addressed in Chapter 6.

Chapter 2

An Overview of the Power Control Techniques for DS/CDMA Cellular Systems

Abstract In this chapter, the cellular systems and various power control schemes are briefly reviewed. Power control is one of the most important factors for DS/CDMA cellular systems due to the near-far problem and the multipath fading.

Without power control, the capacity of the DS/CDMA cellular system could become lower than that of cellular systems based on FDMA. A proper power control will make users to share resources of the system equally between themselves and thus enhance the system capacity to meet the demanding requirements.

2.1 Introduction

Code Division Multiple Access (CDMA) has become the technology of choice for the third generation of cellular mobile systems because theoretically it can provide higher capacity compared with FDMA and TDMA schemes. However, in order to achieve the high capacity, techniques are needed to overcome several wireless impairments. One of the most important key aspects of CDMA is the power control. Many research works have been devoted to investigate the power control of DS/CDMA cellular systems and even design the power control algorithm to achieve an optimal performance. This chapter wills briefly overview the relevant background related to this hot topic in DS/CDMA cellular systems.

The rest of this chapter is organized as follows. An overview of the cellular systems is given in section 2.2. The power control techniques in DS/CDMA cellular systems and their corresponding classifications are presented in section 2.3. Finally, the concluding remarks are given in section 2.4.

2.2 Cellular Systems

The 3G cellular services, known as IMT-2000, will sustain higher data rates and open the door to many Internet style applications. The most important IMT-2000 proposals are the W-CDMA as the successor to GSM, cdma2000 as the successor to IS-95, and time-division synchronous CDMA (TD-SCDMA). In 3G systems, wideband CDMA has been chosen because theoretically it can provide higher capacity compared with FDMA and TDMA schemes. However, in order to achieve this “promised” high capacity, good techniques are needed to overcome several wireless impairments. This is why significant research works are currently being devoted to improve the performance of DS/CDMA systems, such as interference cancellation or multiuser detection, smart antennas, and power control, etc. Among those areas of research, power control is the most crucial aspect because it plays an important role in a DS/CDMA cellular system.

Without good power control scheme, the capacity of a DS/CDMA system may be only comparable with or even less than the capacity of FDMA or TDMA systems. Henceforth, in this dissertation, it is motivated to analyze the performance of the power control schemes in DS/CDMA cellular systems.

2.3 Power Control in DS/CDMA Cellular Systems

The necessity for power control in FDMA/TDMA-based cellular networks stems from the requirement for co-channel interference management. This type of interference is caused by frequency reuse due to limited available frequency spectrum. By a proper power adjustment, the harmful effects of co-channel interference can be reduced. This allows a more “dense” reuse of resources and thus higher capacity.

The reason why the power control is the most important issue for a DS/CDMA cellular system is related to the unique feature of the CDMA system, MSs in the system are interfering with each other. In a DS/CDMA system, many users can transmit messages simultaneously over the same radio channel, each using a specific spread-spectrum pseudo-noise (PN) code [10].

Within a cell, the code channels in downlinks can be considered as mutually orthogonal because downlinks may exhibit synchronous CDMA transmission. However, these code channels in uplinks cannot be exactly mutually orthogonal for a set of asynchronous users, and thus mutual interference occurs among the uplinks. In such a case, a strong signal increases communication quality, and a weak signal may suffer from strong interference. This problem is referred to as the near-far effect and limits the CDMA system capacity [11]. Hence, well-designed power control is essential for proper functioning of the DS/CDMA cellular system. In the absence of power control the capacity of the DS/CDMA cellular system is very low, even lower than that of mobile systems based on FDMA or TDMA. Power control techniques in DS/CDMA systems can be classified in many different ways, which are described in-detail in the following subsections.

2.3.1 Uplink/Downlink Power Control

Power control schemes can be classified into two categories: uplink power control scheme, and downlink power control scheme. The uplink is also known as the reverse link, and the downlink is also known as the forward link. Uplink power control is found the single most important requirement for DS/CDMA systems due to the near-far problem [3]. Without uplink power control, i.e., each MS's transmission power is the same; the system capacity is found to be unacceptably low. In order to achieve the same average received signal power level, it is reported that a dynamic range of around 80dB while adapting MS's transmission power is required [3].

For the downlink link, there is no near-far problem since all signals are transmitted and hence vary together. In a single cell system, no downlink power control is required. However in a cellular system, interference from neighboring cell sites fades independently from the given cell site and thereby degrades performance. Thus it is necessary to apply power control in this case also, to reduce intercell interference. In this dissertation, we will focus on the uplink power control. And the contents in next coming sub-sections are all based on the uplink power control.

2.3.2 Centralized/Distributed Power Control

Power control schemes can be classified into two categories: centralized power control and distributed power control. A centralized power control is assumed to have all information about the gain of each link between any MS and any BS, and accordingly is capable of well controlling all the MSs' transmission power levels such that carrier-to-interference (CIR) balancing is met [31]. The concept of CIR-balancing is firstly introduced by Aein [4], in which all receivers experience the same CIR. As one can see, CIR-balancing implies fair resource usage among the users, therefore the outage probability is reduced and the system capacity is enhanced as compared CIR-unbalanced situation. The centralized power control for achieving CIR-balancing was identified as an eigenvalue problem for positive matrices for satellite systems [4],

FDMA/TDMA cellular systems [8], and DS/CDMA cellular systems [32]. The transmission power allocations are determined by the eigenvector corresponding to the minimum positive eigenvalue of the link-gain matrix. The study of the centralized power control reveals the optimal transmission power allocation to achieve CIR-balancing. However, to obtain the link-gain matrix need to know all of the link gains, which is obviously not practical.

The distributed power control is also known as decentralized power control. A decentralized power control scheme [9], [33], [34] adapts only one MS's power, and the algorithm depends only on local information, such as the measured SIR or the estimated gain of the channel from the specific MS to its serving BS. Due to depending only on local information, the distributed power control is more practical than the centralized one. However, the distributed power control would not perform better than the centralized one, e.g. in the aspect of convergence speed. Lee [35]

proposed an algorithm to adapt each MS's power merely based on the local information, especially focused on the convergence issue of the power adaptation.

All the works studying either the centralized or distributed power control are also assumed rather ideal system model without considering the practical issues such as, (1) the power is practically adapted according to the periodically issued power control commands, (2) a power control command can only contain finite bits, (3) the existence of the power control command delay, (4) the short-term channel fading. Once these practical factors are further considered, the convergence of the distributed power control can become a very complex problem. The study of the centralized/distributed power control can indeed provide the first view to the power control scheme in a DS/CDMA cellular system without getting into the detail of practical issues.

2.3.3 Open-loop/Closed-loop/Outer-loop Power Control

From a viewpoint of realistic system, the power control techniques can be classified into three categories: open-loop power control, closed-loop power control, and outer-loop power

control.

The open-loop power control, also called average power control, is used to overcome the near-far and shadowing effects on the reverse link of a DS/CDMA cellular system. The open-loop power control is designed to ensure that the average received powers from all MSs in the same cell are equal to a preset level. By utilizing that the large-scale propagation loss is substantially reciprocal between uplink and downlink channels, the MS can compute the required transmission power based on the estimated channel gain on the downlink [3]. Such a task can be done by the MS itself without needing any feedback information from its serving BS, therefore this kind of power control is of an open-loop type. Due to the mechanism of the open-loop power control is so simple that only few literature, e.g. [36], [37], can be found.

Closed-loop power control aims at compensating for the received signal fluctuation due to short-term channel fading (or called multipath fading), which cannot be eliminated by the open-loop algorithm since the short-term channel fading can be very difference for the forward and reverse links. Usually, when talking about the closed-loop power control, it has implicitly assumed that the open-loop power control is a perfect one such that the near-far effect is perfected compensated. Due to the short-term fading is uncorrelated between uplink and downlink, the BS needs to measure some quality indicator, e.g. the received signal or the received SIR, and then sends a corresponding power control command back to the MS, so that the MS can adjust its transmission power accordingly. Obviously, this is a control mechanism in a closed-loop manner.

Depending on what kind of quality indicator is measured at the BS, the closed-loop power control is further divided into three major classes. Please refer to section 2.3.4 for further detail.

Typical operation of a closed-loop power control is described as follows. The quality indicator, such as the received signal power, received SIR, or the bit error rate (BER) is estimated or measure at the BS for every power control period, T^p. In order to efficiently compensate for the short-term fading, the power control period should be shorter than the average fade duration of

the short-term fading. Then estimated quality indicator is compared with a preset target threshold.

A finite-bit power control command is generated based on the difference between the estimated quality indicator and the target threshold and sent back to the MS via the downlink channel. The power control command is usually multiplexed with the downlink user data. The MS then extracts the power control command from the downlink data stream and adjust its transmission power by an amount of CMD*∆p (dB) for the next power control period, where CMD denotes the power control command and ∆p is a preset step size in dB. Note that a delay is always introduced by such a control loop. This delay is called the power control loop delay. Due to lots of factors such as the power control period, finite-bit power control command, power control loop delay, etc., each MS's transmission power adaptation can never compensate the channel fading such that the corresponding quality indicator seen by the BS is always kept at the target threshold. In other words, the quality indicator seen by the BS would be a random variable. The quality indicator will have its mean around the target threshold. The difference between the estimated quality indicator and the target threshold is usually called power control error.

Note that closed-loop power control is only effective if and only if the validity of the power control command is good enough. A factor affects the validity of the power control command is the power control loop delay. A too-long power control loop delay will reduce the validity of the power control to reflect the current channel variation. Due to this fact, closed-loop power control is feasible in a terrestrial cellular environment, while not in a mobile satellite communications systems. Another factor is the speed of the multipath fading, or the MS mobility or velocity, which can also reduce the validity of the power control. The study of impact of the MS mobility on the system performance can be found in [38].

The outer-loop power control always cooperates with the SIR-based closed-loop power control, since for a typical strength-based power control, the target threshold is considered fixed and there is no corresponding outer-loop power control. The outer-loop power control is

employed to adapt the target SIR threshold used in the SIR-based power control such that the ultimate goal, the average BER, can be maintained at a certain level [39]. Note that, identical target SIR threshold for communication links does not imply identical BER performance. To speak more specifically, although identical target SIR threshold does imply that the received SIR at BS has its average around the same target SIR threshold, however, the variance of the received SIR is affected not only by the target SIR threshold but also by many other factors such as channel fading conditions, MS velocity, etc. Therefore, identical target SIR threshold can not guarantee identical distribution of the received SIR as well as identical average BER performance.

In realistic system, different MSs may require different target SIR threshold and the outer-loop power control is needed to adaptively adjust the target SIR threshold in order to achieve the target BER performance for each uplink. The algorithm of updating target SIR threshold can be found in [40], [41], [42], [43]. To determine a proper target SIR threshold, the BS should be capable to estimate the average BER. The estimated average BER is then compared with a target BER, which might be difference for different type of communication services. If the estimated average BER is better than the target BER, the target SIR threshold is decreased; otherwise the target SIR threshold is increased. In fact, the outer-loop power control is a kind of BER-based power control.

2.3.4 Strength-based/SIR-based/BER-based Power Control

According to what kind of quality indicator is measured to determine the power control command, the closed-loop power control can be classified into three categories: strength-based, SIR-based, and BER-based power control schemes.

In the strength-based power control scheme, the strength of the signal received by the BS from an MS is measured and compared with a preset target threshold. A one-bit power control command 'UP'/'DOWN' is periodically issued depending on whether the received signal power is lower/higher than the target threshold. In general, the power control command can be of more

than one bit, however in practical system the one-bit power control command is implemented due to its simplicity and light bandwidth requirement. For a typical strength-based power control scheme, the target threshold used in each BS is the same. Such a strength-based power control scheme will never diverge since the received power at the BS as well as the MS's transmission power will always converge. This is an advantage of the strength-based power control scheme.

However, as can be easily found, the unification of the target threshold in the whole system is a negative factor to the system capacity. Typically, the central cells in the system will suffer from much interference than the boundary cells. Therefore the received SIR in the boundary cells will be unnecessary high and therefore the corresponding MSs will induce too much interference to the system. If these target thresholds that are unnecessary high can be somehow lowered down,

However, as can be easily found, the unification of the target threshold in the whole system is a negative factor to the system capacity. Typically, the central cells in the system will suffer from much interference than the boundary cells. Therefore the received SIR in the boundary cells will be unnecessary high and therefore the corresponding MSs will induce too much interference to the system. If these target thresholds that are unnecessary high can be somehow lowered down,