为了提高有限长LDPC码的打孔性能,提出一种码率兼容LDPC码的打孔算法。结合打孔变量点恢复树结构的特点,采用贪婪搜索算法逐级最大化k步可恢复节点的数量,以获得尽可能多的k值较小的k步可恢复节点,从而改善码率兼容系列子码的误码性能。针对随机构造和PEG构造下的LDPC码,验证了本文打孔算法的性能。仿真结果表明,本文方法生成的系列码率兼容子码的误码性能均优于随机打孔方法,,特别是当子码的码率较高时,误码性能改善越明显。
In order to improve puncturing performance of finite length LDPC codes, a punctured algorithm to achieve rate-compatible LDPC codes is proposed in this paper. Considering the characteristic of recovery tree rules of punctured variable nodes, the greedy search algorithm is used to stepwise maximize the number of punching variable nodes of each k-step recoverable,in order to obtain more k-step recoverable nodes when k is small, thereby improving the BER performance of rate-compatible sub-code. Randomly constructed LDPC codes and PEG algorithm constructed LDPC codes verify the puncturing performance. The experiment results show that optimized puncturing is superior to randomly punctured method over wide range of rates. Espe- cially when the code rate is high, the bit error performance improvement is more obvious.