基于MMSE的大气激光通信中LDPC码BP译码算法研究

Optimization of Two BP-based Decoding Algorithms for LDPC Codes Based on MMSE Criterion under Weak Turbulence Channel

针对大气激光通信中低密度奇偶校验码(LDPC)置信传播(BP)译码算法复杂度高的问题,对几类BP-Based简化译码算法进行了分析,并基于最小均方误差准则(MMSE)对Scaled BP-Based和Offset BP-Based两类改进译码算法的优化设计进行了探讨,得出两类算法的最优校正因子,并给出了数值计算.在不同的湍流强度下,对码长1008的(6,3)比特填充LDPC码进行了仿真实验,结果表明,对于短码长的LDPC码,当译码BER=10-5时,最小和算法(UMP BP-Based)相对于BP算法有0.1～0.15dB的译码性能损失;基于MMSE设计的两类算法相比BP算法大大降低复杂度的同时,译码性能与BP算法相当,甚至优于BP算法,优于UMP BP-Based算法0.075～0.15dB.

Due to high complexity of Belief Propagation(BP) algorithm,several kind of BP-based simplified decoding algorithms are analyzed.Based on the Minimum Mean Square Erro(MMSE) criterion,optimal design of the scaled BP-based and offset BP-based algorithms is discussed.The theoretical formulas and numerical calculations on the optimal factors for these two BP-based algorithms are provdied.The simualtion resutls for the(6,3)regualr LDPC codes of lengths 1008 on the weak atmospheric turbulence channel demonstrate that the scaeld BP-based and offset BP-based aglorithms with the proposed factors are better than Min-Sum Algorithm,and even can achieve the performance better than that of the BP algorithm,and have lower complexity compared with BP algortihm.