中央控制架構下裝置到裝置通訊網路的資源分配

Abstract

在蜂巢式網路(cellular networks)環境下，裝置對裝置通訊(Device-to-Device communication, D2D)將成為未來無線網路一項重要的技術。D2D 通訊產生新的網路服務機會，並且降低基地台對於短距離通訊的負載，也藉由空間重複增益的特性，來提升頻譜使用效率。本論文考慮在中央控制的D2D通訊架構下，蜂巢式網路營運商對直接通訊移動裝置進行無線通訊資源排程，以同時達到最大化系統效能及資源公平性分配的目標。本論文所提出的資源排程演算法將直接通訊資源劃分為資源專用區間和資源競爭區間，並考慮空間重複增益和滿足比例式公平性來做為決定準則，進行動態調整資源分配區域和資源競爭區域比例。我們提出兩種易於實現資源排程演算法：演算法一主要考量是否所有D2D使用者達到理想的資料傳輸速率，進而決定資源區域邊界；而演算法二基於Proportional Fair (PF) metric值，來決定資源區域邊界。模擬結果顯示，本論文所提出的兩種演算法皆可有效利用空間重複增益來提高系統整體效能，並且維持所有D2D使用者擁有資源比例式公平性。

Device-to-Device (D2D) communications underlying cellular networks potentially will become an important technology in future networks. D2D communication enables new services and reduces eNB load for short range peer-to-peer communications. It also can improve spectrum efficiency because of resource spatial reuse. In this thesis, we consider a central-controlled D2D architecture, in which a cellular operator can decide whether mobile devices can connect to each other directly. We aim to maximize system throughput while also achieve fairness in resource allocation. A resource scheduling algorithm that dynamically adjusts the boundary of resource-assignment area and resource-contention area is proposed. Also, criteria to extract spatial reuse gain and meet Proportional Fair (PF) are determined. We propose two resource scheduling algorithms that are easy-to-implement. The first algorithm mainly considers whether all D2D users achieve average ideal data rate to determine the resource area boundary. The second algorithm is based on PF metric to determine the resource area boundary. Simulation results show that the two proposed algorithms can increase system throughput by exploring spatial reuse gain and maintain good proportional fairness among D2D communication users.