用於多用戶偵測之適應性盲蔽平行式干擾消除接收機

Abstract

直接序列碼分進接 (DS-CDMA) 是第三代行動通訊中標準的多重進接技術。眾所週知， 此通訊技術之系統效能常受限於多用戶干擾 (MAI)，因此有多種多用戶偵測 (MUD) 技 術用來改善此問題。平行式干擾消除法 (PIC) 藉由其低複雜度與優良表現之故，已成為 一深具潛力的改善方案。再者，藉著引進部分消除因子 (PCF)，一PIC 的改良架構也在 發展中，稱之為部分平行式干擾消除法。雖然最佳部分消除因子可用於增加系統效能， 但其最佳值並不易求得。在本計畫中我們提出一新型之適應性濾波器來克服此難題。我 們設計一具濾波效果之接收機，藉由最小均方差之適應性理論求得最佳部分消除因子。 此法之最大優點在於不需要訓練序列且其複雜度很低。初步的實驗結果顯示，我們提出 之方法可大幅提升傳統適應性部分平行式干擾消除器之系統效能。

Direct-sequence code-division multiple-access (DS-CDMA) has been adopted as the standard multi-access technique for the third generation wireless communications. It is well known that he performance a DS-CDMA system is limited by multiple-access interference (MAI). Many multiuser detection (MUD) algorithms were proposed to alleviate this problem. Due to its simplicity and good performance, parallel interference cancellation (PIC) has been considered a promising MUD algorithm. By introducing partial cancellation factors (PCFs), a variant of PIC called partial PIC was also developed. Although the partial PIC can have improved performance, the optimal PCFs are difficult to derive. In this project, a new type of adaptive filtering approach is proposed to overcome this problem. Special designed filtering architectures are developed and optimal PCFs are trained using the least mean square (LMS) adaptive algorithm. The distinct features of the approach are that it does not require any training sequences and the computational complexity is low. Primitive simulation results show that the proposed algorithm can significantly outperform the conventional adaptive partial PIC approach