高速分碼多工存取系統之軟式交遞及下行功率分配性能分析研究

Abstract

在這篇論文中，我們研究了軟式交遞(Soft Handoff)和功率分配(Power Allocation)機制在寬頻分碼多工存取系統(Wide-Band Code Division Multiple Access, WCDMA)中的效能分析。在此論文中，我們提出了一套等高線重疊法(Contour Overlapping)來優化寬頻分碼多工存取系統的軟式交遞參數。除此之外，我們針對不同的功率分配機制從不同角度進行評量。我們從以下三個方向評量功率分配機制：1. 功率分配效能(Power Efficiency)； 2. 對功率控制錯誤(Power Control Errors)的敏感度； 3. 對頻道快速變動衰減(Channel Fast Fading)的抵抗能力。本篇論文主要分成三個部份。

首先，我們提出了一套利用等高線重疊法來優化寬頻分碼多工存取系統的軟式交遞參數的方法。優化軟式交遞參數的優化是一個相當具有挑戰性的研究題材，因為在進行優化的過程中，必須同時考量到數個系統效能需求。在本論文中，我們考慮了四個主要的軟式交遞參數，其中包括 AS_TH、AS_TH_HYSY、AS_REP_HYST和AS_MAX_SIZE。不同的參數設定會影響從以下幾個層面影響系統效能： 1. 執行軟式交遞的機率； 2. 軟式交遞連線個數(Active Set Size)； 3. 連線品質(Link Quality)。從以上三個系統效能評量，我們針對每個參數進行單獨的評量。由個別評量的結果發現，並不是所有的參數對於系統效能都有顯著的影響，例如AS_REP_HYST和AS_MAX_SIZE這兩個參數的影響就很有限。根據以上觀察，我們針對另外兩個參數(AS_TH和AS_TH_HYSY)設計出一套優化參數流程，以期同時滿足數個系統效能要求。

第二部份中，我們利用數學分析在不同功率分配機制下，因為功率效能不張以及對功率控制錯誤所造成通訊中斷(Outage)的機率。功率分配機制主要是在進行軟式交遞的時候，分配同時連線中基地台的傳送功率，以期達到要求的雜訊比(Signal to Interference Ratio)。在本論文中我們分析四種不同的功率分配機制，其中包括品質平衡工率分配法(Quality Balancing Power Allocation, QBPA)、等能量功率分配法(Equal Power Allocation, EPA)、基地台多樣化選擇分配法(Site Selection Diversity Transmission, SSDT)和連線比例式分配法(Link Proportional Power Allocation, LPPA)。在同時考慮基地台功率不足和功率控制錯誤的影響下，我們提出了一個評量中斷的計算方式。從我們的結果顯示，儘管SSDT是最有效率的分配法則，但是同時也因為對功率控制錯誤比較敏感，其中斷機率在一般的系統負載下相對比較高。反之LPPA不但幾乎可以達到SSDT的系統容量，同時也可以維持相對較低的中斷機率。 最後，我們提出了一個分析方法用來計算在不同功率分配機制下因為通道快速衰減的影響下所造成的通訊中斷機率；以及計算預估使用者在不同功率分配機制下所消耗的能量。在此分析中，我們考慮了萊斯衰減(Rician Fading)通道並且考慮了同頻段干擾(Co-channel Interference)。根據計算結果，SSDT所送出的功率最低，驗證了我們在前一個部份的証明。再者，從對快速衰減的抵抗力來看，SSDT最差而QBPA最好。比較SSDT和LPPA，我們發現LPPA藉由損失些許功率效率，提供較高對快速衰減的抵抗力。 綜合整篇論文重點，主要進行了軟式交遞和功率分配機制在寬頻分碼多工存取系統中的效能分析，預期藉由選擇適當的系統參數以及能量分配機制，提升系統整體的效能。

In this thesis, we investigate the soft handoff mechanism and power allocation mechanism for the wide-band code division multiple access (WCDMA) system. We propose a contour overlapping technique for soft handoff parameters design in the WCDMA system. Moreover, we evaluate the performance of different power allocation schemes subject to power efficiency, sensitivity to power control errors and resistance to the fast fading. This thesis includes three parts.

First, we presents a contour overlapping technique to design downlink soft handoff parameters for the WCDMA system. Optimizing soft handoff performance in the WCDMA system is a challenging task because it is necessary to determine many parameters to meet different requirements simultaneously. There are four key soft handoff parameters, including AS\_TH, AS\_TH\_HYST, AS\_TH\_REP, and AS\_MAX\_SIZE. Each soft handoff parameter affects different system performance aspects, including handoff probability, active set size, and link quality. Through simulations, we evaluate the impact of each handoff parameter on different performance metrics. Our results indicate that AS\_TH and AS\_TH\_HYST significantly affect the target $E_b/N_o$, handoff probability, and average active set size, while the other two parameters, AS\_TH\_REP and AS\_MAX\_SIZE, have little effect. Based on this observation, we develop a simple contour overlapping technique to determine handoff parameters AS\_TH and AS\_TH\_HYST subject to various performance requirements simultaneously. The proposed design method can facilitate the soft handoff parameters design for the WCDMA system.

Second, we derive the closed-form expressions for the outage probabilities of different power allocation mechanisms subject to transmission power control errors. In addition to site selection diversity transmission (SSDT), we consider the equal power allocation (EPA), quality balancing power allocation (QBPA), and link proportional power allocation (LPPA) schemes. From both the perspectives of base station power shortage and power control errors, we develop a semi-analytical approach to evaluate the outage performance of different power allocation schemes. Our numerical results show that due to power control errors, the system with SSDT suffers from a higher outage probability with a moderate traffic load although it can achieve the lowest outage probability performance in the heavy traffic condition. Furthermore, the LPPA system can almost approach the same capacity as SSDT system, while maintaining lower outage probability than SSDT in the typical traffic condition.

Third, we present an analytical approach to evaluate the outage probability of a handoff user in the WCDMA system when applying different downlink power allocation mechanisms in the log-normally shadowed and Rician fading channel subject to co-channel interference. We consider the following power allocation mechanisms, including SSDT, LPPA, EPA and QBPA. An analytical formula for calculating the power efficiency of the above schemes is also derived in the paper. Out results show that although the SSDT method is most power efficient, it is also most sensitive to the fast fading and co-channel interference. As compared to the SSDT method, the LPPA mechanism achieves better outage in presence of fast fading at the cost of slightly sacrificing power efficiency. The proposed analytical frame work can also evaluate the effects of soft handoff threshold, orthogonality factor of channelization codes, and the Rician factor.

To summarize, we analyze the soft handoff and power allocation mechanism in the WCDMA system. For soft handoff study, we provide a systematic methodology to design the optimal soft handoff parameters subject to multiple system requirements. For power allocation schemes study, we evaluate the performance of different power allocation schemes, including EPA, QBPA, SSDT, and LPPA, from the perspective of power efficiency, resistance to power control errors and sensitivity to Rician fast fading effects considering co-channel interference. Through the study of this thesis, we provide some guidelines to improve the system performance by adjusting soft handoff parameters and power allocation schemes, which can be implied to the real systems at least cost.