基于并行分量译码的分段拟合Turbo码译码

A Turbo Decoding Method based on Parallel-Concatenated Decoder and Piecewise Fitting

为减小迭代译码和Log-MAP算法因运算复杂度高而引起的译码延迟,同时保证其低误码率性能,提出了一种基于并行分量译码的分段拟合Turbo码译码方案。该方案首先改变了传统分量译码器的串行结构,两个分量译码器并行译码;其次利用最小二乘法分段线性拟合Log-MAP算法中的相关函数,能有效降低译码延迟和运算复杂度。最后仿真表明新方案与传统译码相比有一致的误码率性能,且迭代时间缩短了一半。

To reduce the turbo decoding delay caused by iterative decoding and the high complexity of Log-MAP algorithm while keeping the low error-rate performance,this paper proposes a turbo decoding algorithm based on parallel-concatenated decoder and piecewise fitting.Firstly,the algorithm decodes with the two parallel-concatenated decoders.Then by using the least squares method to fit the correlation function of Log-MAP algorithm,the decoding delay and computational complexity could thus be reduced.And finally,the simulation shows that the new scheme,as compared with the traditional decoding scheme has the consistent error-rate performance,and its iteration time is reduced by half.