低密度奇偶校验码快速收敛译码算法研究

Research on a Fast Convergence Decoding Algorithm for LDPC Codes

介绍低密度奇偶校验码(LDPC码)的构造方法和置信传播译码算法,引入基于校验节点的一种快速收敛译码算法——串行译码算法。从树的深度方面分析串行译码算法的消息收敛特性,证明该算法与置信传播译码算法相比具有较好的收敛特性,且降低了译码复杂度。在加性高斯白噪声(AWGN)环境下,采用BPSK调制方式分别对串行译码算法和置信传播算法进行了计算机仿真。结果表明,串行译码算法的译码性能具有明显的改善。该算法使硬件实现变得更容易,资源占有量会降低,这就为LDPC码的工程实现提供了一种可行的方案。

The construction methods of LDPC codes with its Belief Propagation(BP) decoding algorithm are briefly introduced.On this condition,a fast convergence decoding algorithm based on check nodes is introduced,which is called serial decoding algorithm.By analyzing the tree depth of the message convergence,it is proved that the convergence speed of this algorithm is better than the BP decoding algorithm and the complexity of this algorithm is reduced.Then,the BP decoding algorithm and the serial decoding algorithm are simulated respectively for BPSK signals in the AWGN channel.The results of the simulation show that the decoding performance can be improved by the serial decoding algorithm.The serial decoding algorithm is simple.Its hardware implementation is easier and it needs less resources.So the serial decoding algorithm is proved to be an effective method of Low Density Parity Check codes in the engineering implementation.