摘要
极化码的置信传播(Belief Propagation,BP)译码算法性能相比于其他极化码译码算法并不具有优势。为了改善这一现象,提出了一种基于残差网络和扰动译码算法相结合的BP译码算法。该算法通过在传统BP译码算法的基础上添加残差神经网络对接收信号进行处理,使其更大概率地落在可正确译码区域内,从而达到改善传统BP译码算法的译码性能的目的。仿真结果表明,在误比特率为10时,所提算法相比于传统的BP译码算法约有0.7 dB的性能增益,相比于BP-RNND(50)(BP-Residual Neural Network Decoder)译码算法约有0.6 dB的性能增益;同时,在低信噪比时所提算法的平均迭代次数相比于传统BP译码算法约有60%的降低。
The performance of Belief Propagation(BP)decoding algorithm is not superior over that of other polar code decoding algorithms.In order to improve its performance,a BP decoding algorithm based on residual network and perturbation decoding algorithm is proposed.By adding residual neural network to the traditional BP decoding algorithm to process the received signal,the algorithm makes it fall in the correct decoding area with greater probability,so as to improve the decoding performance of the traditional BP decoding algorithm.Simulation results show that the performance gain of the proposed algorithm is about 0.7 dB compared with that of the traditional BP decoding algorithm and 0.6 dB compared with that of the BP-Residual Neural Network Decoder(BP-RNND)(50)decoding algorithm at bit error rate of 10.At the same time,the average number of iterations of the proposed algorithm is about 60%lower than that of the traditional BP decoding algorithm at low signal-to-noise ratio.
作者
王华华
徐勇军
秦红
方泽圣
WANG Huahua;XU Yongjun;QIN Hong;FANG Zesheng(School of Communication and Information Engineering,Chongqing University of Posts and Telecommunications,Chongqing 400065,China)
出处
《电讯技术》
北大核心
2022年第8期1161-1165,共5页
Telecommunication Engineering
基金
教育部-中国移动科研基金(MCM201805-2)。
关键词
极化码
置信传播译码算法
残差网络
扰动噪声
polar code
belief propagation decoding algorithm
residual network
perturbation noise
