改进的基于剩余度置信度传播的LDPC解码算法

Decoding for LDPC codes with enhanced residual belief-propagation

低密度奇偶校验(LDPC,Low-Density Parity-Check)码的剩余度置信度传播(RBP,Residual Belief-Propagation)和基于行的剩余度置信度传播(NWRBP,Node-Wise RBP)解码算法的性能提升非常有限且计算复杂度较高.提出改进的RBP(ERBP,Enhanced RBP)算法,在一个子迭代中,仅更新一个消息,然后设置被更新消息所在行的所有节点的剩余度值为0,使得ERBP解码算法在每个子迭代中使用不同行的消息进行计算,以加速迭代收敛.不同的LDPC码用于对所提出的算法进行性能仿真.仿真结果表明,与其他算法相比,ERBP算法降低了误帧率(FER,Frame Error Ratio),并加快了迭代收敛速度.

The performance improvements of residual belief-propagation （RBP） and node-wise RBP （NWRBP） decoding algorithms for low-density parity-check （LDPC） codes are very limited while at cost of high computational complexity. The enhanced RBP （ERBP） decoding algorithm was proposed, which updates only one message and then sets the residuals of all messages in the row of the updated message to 0 in one sub- iteration, thus the ERBP can utilize the messages in different check equations in each sub-iteration to increase the iteration convergence rate. Different LDPC codes were used to test the performance of the proposed algo-rithm. The simulation results show that the proposed algorithm, when compared with other algorithms, lowers frame error ratio （FER） and speeds up the iterative convergence.