通道匹配疊代演算法之研究

Abstract

在數位通訊系統，訊源編碼器的設計受限於複雜度與延遲，其編碼輸出序列仍有殘餘冗息存在。為了加強系統的強健性，我們在接收端以渦輪碼原理設計疊代解碼演算法。其關鍵在於利用訊源殘餘冗息和通道編碼提供的保護位元，在通道解碼器與訊源解碼器之間交叉利用解碼輸出的額外訊息。藉由疊代次數增加，額外訊息的準確性也跟著增加，而解碼效能也隨之提升。此外，在無線通道環境中常出現叢集性錯誤，疊代解碼設計應考量錯誤之間的相關性。模擬結果顯示，基於迴旋碼的疊代解碼效能較好，特別是與通道匹配的解碼機制可進一步改善系統強健性。

Due to complexity and delay limit of source encoder, there exits source residual redundancy in digital communication. In order to enhance the system robustness, we proceed with iterative source-channel decoding based on a turbo principal at the receiver. The basic strategy is to exchange extrinsic information can be between the channel decoder and source decoder. Real channel are characterized by error burst due to the combined effects of intersymbol interference and multipath fading. It is therefore believed that further enhancement can be realized through a more precise characterization of the channel on which the iterative decoder design is based. Simulation results indicate that channel-matched iterative decoding based on convolutional code achieves the highest robustness against channel errors.