一种改进的最大似然软判决译码

An Improvement to Maximum-likelihood Soft Decision Decoding

本文提出了一个基于Dijkstra’s算法 (DA)的线性分组码最大似然软判决译码 (SDA) ,与已有的译码方法相比 ,SDA译码具有新的特点 :( 1 )采用新的度量函数 ,使计算更简单 ;( 2 )采用更有效的搜索算法———DA算法 ,实现最大似然软判决译码 ;( 3 )建立错误图样的广义门限 ,进一步加快译码速度 .模拟表明 ,与其它软判决译码算法相比 ,该译码算法在保持最优译码性能的同时 ,能明显地提高译码速度 .同时指出 :采用非最佳信号形式会导致性能损失近3dB .

An improved maximum likelihood soft decision decoding (SDA),using the Dijkstra's algorithm,is presented.Compared with the decodings available,the proposed decoding algorithm has following new advantages:(1)using a new metric function,which reduces decoding calculation;(2)applying a more efficient search algorithm,called DA to the maximum likelihood soft decision decoding;(3)setting up a new general threshold for error patterns to speed up the decoding further.Our simulation results show that SDA is much faster in mean decoding speed than other soft decision decoding algorithms without any loss of decoding performance.Meanwhile,it is emphasized that usage of non optimal transmission signal will result in nearly 3dB loss of decoding performance.