本論文發展基於基因演算法之多用戶偵測器(Genetic Algorithm-based Multi-User Detector, GA-MUD),以改善多載波分碼多工存取(Multi-Carrier Code Division Multiple Access, MC-CDMA)行動通訊系統之效能與運算複雜度。由於傳統GA-MUD在高用戶數時,無法以有限的代數(Generation)大幅改善系統效能,因此我們提出一種修正型GA-MUD架構,此架構將強迫突變(Forced Mutation, FM)加入每代演化中。於選擇性頻率衰減(Frequency Selective Fading)通道下的模擬結果顯示,FM-GA-MUD可在極低運算複雜度的情況下獲得與傳統GA-MUD相同的效能。除外,實驗結果亦顯示,FM-GA-MUD加上傳統之平行干擾消除(Parallel Interference Cancellation)技術,可進一步大幅降低位元錯誤率(Bit Error Rate, BER)。
In this study a novel genetic algorithm based multiuser detector (GA-MUD) is developed to improve the performance and computational complexity of the multi-carrier code-division multiple-access (MC-CDMA) communication systems. To overcome the difficulty of high computational complexity of the conventional GA-MUD, a forced mutation (FM) operation is proposed and employed to the initialization procedure at each GA generation. A series of simulation is considered over frequency selective fading channel and the simulation result indicates that our proposed scheme can achieve the same BER of the conventional GA-MUD with a considerable reduction in computation. Furthermore, this study also finds that the bit error rate can be improved substantially when partial parallel interference cancellation (PPIC) is applied to the proposed FM-GA-MUD.