基于树状图的不等保护分组码译码算法

Decoding Algorithm of Unequal Error Protection Block Code Based on Tree Topology

提出一种新的不等保护分组码译码算法——树状图算法.该算法通过对生成矩阵的初等变换,把码字分解成数个码元组,用这些码元组构成树状图的分支字,并使树状图的每一级对应一位信息元,然后在该树状图上搜索最大似然码字,并由此译码.对于信息序列中所有保护能力大于或等于码字中错误比特个数的信息元,该算法都能保证其准确译出,并且大幅度降低了不等保护码译码的运算量,实现了快速译码.

This paper presents a new decoding algorithm of unequal error protection （UEP） code, which is called tree topology. The algorithm divides the codeword into several groups by transforming generator matrix. These code groups form branch words of the tree and every level in the tree denotes an information bit. Then decoding rules are used to search for the maximum likelihood codeword in the tree topology. This algorithm can correctly decode all information bits with protection capability equal to or higher than the number of error bits in a codeword. Furthermore, this algorithm simplifies the operation and achieves quick decoding of UEP code.