行動通訊網路之無線擷取控制與資源管理

Abstract

由於無線界面是連接行動通訊系統與行動用戶的唯一管道，因此無線擷取控制與資源管理成為行動通訊中最重要的課題。在本論文中，我們探討相關的無線擷取控制與資源管理策略，包括功率控制、多媒體無線擷取控制、以及多媒體資源允諾與分配。 我們首先探討DS/CDMA蜂巢式行動無線通訊系統中的向下鏈路功率控制方法，我們提出了兩種功率控制法：一種是集中式訊號干擾比平衡功率控制法，另一種則是分散式訊號干擾比功率控制法。我們建構了一個電腦模擬的環境來模擬DS/CDMA系統的訊號傳播現象，包括遠近效應、多路徑短期衰減效應、以及自我干擾性質等，以探討這兩種向下鏈路功率控制方法的系統效能。我們藉此獲得這兩種功率控制方法的最佳系統參數，並對他們的損壞機率、穩定性、以及運作難易度作廣泛的比較。 接著我們針對採用動態TDMA協定的多媒體通訊系統，提出一個可保障服務品質的乏晰傳輸控制器。在動態TDMA協定中，使用了小型時槽（mini-slot）的技術來降低隨機擷取機制的競爭成本，並且將這些小型時槽再區分成幾個媒體擷取控制區，分別供具備不同服務品質需求的用戶來使用。我們成功地應用了乏晰邏輯控制理論，包含Sugeno的位置方向型推論法則與最大最小推論法則，來適當地調整媒體擷取控制區與擷取機率，在維持服務品質的條件下增進動態TDMA協定的效率。電腦模擬的結果顯示，我們所提出的乏晰傳輸控制器確實可以使動態TDMA協定達到保障服務品質需求、 較高的系統使用率、以及較低的封包延遲等。 最後我們探討WCDMA通訊系統中，對於無線資源控制的多速率傳輸控制問題。我們採用一種名為Q-learning的即時加強型學習法則，來達成WCDMA通訊系統中最佳的多速率傳輸控制，以取代較早提出的干擾值傳輸控制法。結果顯示，在WCDMA通訊系統中，使用Q-learning的多速率傳輸控制法可以比干擾值傳輸控制法，提供更高的系統容量，並且維持封包錯誤率的服務品質需求。

The radio access control and resource management are most important issues in mobile communications because the air interface is the unique channel to connect a mobile communication and mobile users. In this dissertation, we study the radio access control and resource management schemes, including power control, multimedia radio access control, and multimedia resource admission/allocation. We first investigate the downlink power control methods for a DS/CDMA cellular mobile radio system. Two downlink power control methods are proposed: one is the centralized SIR-balancing power control method and the other is distributed SIR-based power control method. We construct a simulation environment for the propagation effects in DS/CDMA system, including near-far problem, multipath short-term fading, and self-jamming property, to investigate the performance of the two downlink power control methods. The optimal parameters are obtained for the two power control methods and we make comparisons between them in the outage probability, robustness, and operation. And then, we propose a QoS-guaranteed fuzzy transmission controller (FTC) for the the dynamic TDMA protocol in multimedia communication systems. The mini-slot technique is adopted in the dynamic TDMA protocol to reduce contention cost and these mini-slots are further partitioned into multiple MAC regions for different users with respect to heterogeneous quality-of-service (QoS) requirements. We successfully apply fuzzy logic control theory, including Sugeno's position-gradient type reasoning method and the max-min inference method to properly adjust the MAC regions and the access probabilities for enhancing dynamic TDMA efficiency under QoS constraint. Simulation results show that the proposed FTC can indeed help the dynamic TDMA protocol to guarantee QoS requirement and achieve higher system utilization and less non-real-time packet delay. Finally, we investigate the multi-rate transmission control for radio resource control (RRC) in WCDMA communication systems. We employ a real-time reinforcement learning algorithm, named Q-learning, to perform the optimal multi-rate transmission control for WCDMA communication systems, instead of the conventional interference-based transmission control method. The results show that the multi-rate transmission control with Q-learning can improve the throughput of the WCDMA communication system and still keep the QoS requirement of packet error probability guaranteed.