一种低复杂度咬尾卷积码译码算法﹡

摘要

对于咬尾卷积码的译码，传统的最大似然译码算法需要遍历每个可能的起始状态对应的咬尾格形子图，译码复杂度过高。循环维特比算法是一种有效的低复杂度次优译码算法。通过对循环维特比算法中的循环陷阱进行研究，提出了一种新的循环陷阱检测方法，利用对循环陷阱的检测可以减少冗余迭代；同时利用最大似然咬尾路径对非似然起始状态进行排除，极大的缩小了循环维特比算法中译码搜索空间。在此基础上得到了一种低复杂度的译码算法。%It is quite complicated for conventional maximum likelihood decoding algorithm to decode tail-biting convolutional code, for all sub-tail-biting trellises need to be visited. Circular Viterbi algorithm is an effective and low-complexity decoding algorithm. And by studying the circular trap in circular Viterbi algorithm (CVA), a new detection rule for circular trap is proposed and the redundant iteration removed. Meanwhile, the searching space for CVA is reduced by excluding un-likelihood starting states. Based on these, a new CVA-based decoding algorithm for tail-biting convolutional codes is concluded. Compared with other existing algorithms, the new algorithm proposed in this paper could fairly improve the decoding efficiency and reduce the decoding complexity.