Abstract: In order to reduce the bit error rate of message transmission in strong interference channel of power line, a modified algorithm based on belief propagation algorithm for repeat accumulate code is proposed, which through revising the initial value of the information nodes, making use of adaptive and iterative decoding to reduce the number of iterations, and pre-judging whether decoding operation is finished, thus reducing the decoding complexity and the decoding efficiency is improved, too. The simulation under Middleton class-A model shows that the BP algorithm after modification has a performance close to the BP decoding algorithm in the area of low SNR, but in the area of high SNR, the modified algorithm has about 0.8 dB gain than BP algorithm, besides, the iteration times in the decoding process are notably decreased.

Key words: Power Line Communication（PLC）, adaptive and iterative algorithm, belief propagation algorithm, Middleton class-A model, repeat accumulate code

摘要： 为降低信息在强干扰电力线信道中传输的误码率，对重复累积（RA）码的置信传播（BP）译码算法进行改进。修正BP译码算法中信息节点的初始值，利用自适应迭代译码减少译码过程中的迭代次数，并对译码进行预判决，降低了RA码译码复杂度、提高了译码效率。在Middleton A类电力线信道模型下仿真结果表明：改进后的译码方法在低信噪比区域与BP算法性能相近，而在较高信噪比区域改进后的算法与BP译码算法相比所需信噪比少约0.8 dB，并且显著减少了译码过程中的迭代次数。

关键词: 电力线载波通信, 自适应迭代译码, 置信传播算法, Middleton A类信道, 重复累积码