利用软解调序列的LDPC码闭集识别方法

为实现对包括低密度奇偶校验（LDPC）码在内的一系列“稀疏几何编码”的识别，基于软解调序列，设计和实现了一种闭集集合应用背景下的识别方法。以校验约束关系作为切入点，首先回顾了基于硬解调序列的传统识别策略；接着分析和推导了软解调序列所应满足的校验约束形式；通过引入后验校验对数似然比（CLLR）这一统计量,最终设计了相应识别算法。在事先给定先验编码集合的闭集应用模式下，仿真结果显示，算法明显占优，识别增益在低信噪比环境下可达2~6 dB。

A finite set recognition algorithm of LDPC coding by using soft-demodulation sequence

To realize recognition of the sparse-graph coding including low density parity check(LDPC) codes,a finite set algorithm based on the soft-demodulation sequence is proposed.Through making coding parity-relationship as the key point,this paper first reviews the traditional recognition strategy based on hard-demodulation sequence.Then,it analyzes the coding check relationship which soft-demodulation sequence should satisfy.Finally,it designs a new recognition algorithm by introducing a statistics index called posterior check-log-likelihood-ratio(CLLR).The result of simulations which focalize around finite set shows that the new recognition algorithm based on soft-demodulation sequence is better than traditional method based on hard-demodulation sequence,with a recognition gain of 2~6 dB,particularly in low signal-to-noise ratio(SNR).