基于密度进化的LDPC码BP译码算法的量化门限分析

Quantization threshold analysis of BP decoding algorithm for LDPC codes based on density evolution

针对LDPC码BP译码算法中的量化门限的取值问题,提出一种基于密度进化算法的分析方法。该方法从数学角度跟踪迭代译码过程中消息的概率密度函数,并通过计算不同量化门限对应的误码性能来选取最佳量化门限值。仿真结果表明,在分析以量化门限为参数的LDPC码集合时,误码性能会随着量化门限值的提高而逐渐收敛,同时验证了密度进化方法分析量化门限具有高效性、可行性和实用性。

Aiming at the value of the quantitation threshold in BP decoding algorithm for LDPC codes, a new method based on density evolution is proposed. This method tracks the probability density function of the message in the iterative decoding process by mathematics, so as to select the optimal quantization threshold by calculating the error performance of the differ- ent quantization thresholds. Simulation results show that on the analysis of the LDPC codes set which pass quantization threshold as parameter, the BER performance will gradually converge as the increase of the quantization threshold value. And the density evolution method is efficient, feasible and practical.