Research On Successive Cancellation List Decoding Algorithm Of Polar Codes

With the rapid development of the 5G communication technology,polar codes play an important role in improving the efficiency and performance of the communication system.The successive cancellation(SC)decoding algorithm has low complexity and good performance under long codes length.However,the performance of polar codes under short codes length is not good enough.In order to improve the performance of SC decoding,the successive cancellation flip(SCF)decoding algorithm performs multiple decoding attempts while bringing a large decoding delay.The successive cancellation list(SCL)decoding algorithm simultaneously decodes multiple paths,which greatly improves the performance of SC decoding.However,due to the influence of code length and code rate,the complexity and delay of SCL decoding still need to be improved.In order to reduce the complexity of SCL decoding,this paper proposes a new path splitting decision strategy to reduce the number of path splitting(PSN)and sorting operations of path metrics in SCL decoding.The path splitting decision strategy is based on a new auxiliary path metric proposed in this paper.Once the auxiliary path metric exceeds the threshold,the decoding path is no longer considered as the correct path,and the path splitting operation is not performed.If auxiliary path metrics of all surviving paths exceed the threshold,the sorting operation of the path metrics can be avoided.The simulation results show that when N=1024,R=0.5,L=16,and the signal-to-noise ratio(SNR)is 2.25 dB,the number of path splitting of the proposed path splitting decision-aided SCL(PSD-SCL)decoding algorithm is reduced by25.1%,compared with other improved path splitting algorithms,and the performance is almost not degraded.In order to enable PSD-SCL decoding to be used in fading channels,this paper transforms the given fading channel to an equivalent Gaussian channel with the same average Wasserstein distance,which can reflect the geometric characteristics of the channel.The Gaussian channel with the smallest Wasserstein distance from the fading channel is found as an approximate channel model for selecting information channels and decoding.Simulation results show that under fading channel,selecting information channels from the approximate channel model has better performance than the traditional Monte Carlo method.Moreover,when L=2,N=512,R=0.5,SNR=2dB,the PSD-SCL GA algorithm proposed in this paper greatly reduces the decoding complexity with less performance loss and the PSN in Rayleigh fading channels is reduced by 72.9% compared with other path splitting algorithms.