一种简单的Turbo码的迭代停止判据

Turbo码提出之后,由于其优异的性能成为研究热点。但为了得到其优异的性能,在译码过程中需要进行多次迭代,这造成巨大的译码延时。在不影响性能的情况下,为减少译码延时,迭代停止在Turbo码迭代译码过程中十分重要。提出一种基于似然比绝对值大于某一门限的比特数目的迭代停止判据。并且给出平均迭代次数、译码性能的仿真。仿真结果证明该算法的有效性。

A simple iteration stopping criterion for Turbo code

Turbo code becomes focus in coding research because of its excellent performance. But it needs many times of iteration in decoding, which results in long decoding time delay. To shorten the delay, iteration stopping is very important for turbo decoding without performance degradation. This paper proposes a simple iteration stopping criterion based on the number of bits whose likelihood ratio absolute values are larger than certain threshold, and gives simulations of average iteration times and decoding performance. The results of the simulations prove the algorithm’s validity.