Research On Adaptive Successive Cancellation List Decoding Algorithms Of Polar Codes

The polar codes proposed by Professor E.Arikan of Turkey in 2009 is an important coding scheme in the 5G system.The theoretical results show that the polar codes can reach the Shannon limit under the Binary-Discrete Memoryless Channel(B-DMC).However,under the condition of short and medium code lengths,the channel polarization is not complete,resulting in a decrease in decoding performance.This thesis mainly focuses on the research on the ADaptive Successive Cancellation List(AD-SCL)decoding algorithms for polar codes.The adaptive algorithms can reduces the decoding complexity by increasing the number of lists and tentative decoding.However,under the condition of low and medium Signal-to-Noise Ratio(SNR),the performance of Cyclic redundancy check-Aided Successive Cancellation List(CA-SCL)decoding algorithms of different list sizes is degraded,resulting in a sharp increase in the decoding complexity of adaptive algorithms.In order to improve the performance of CA-SCL decoding algorithms,this thesis designed an adaptive list method based on a new type of parity-check concatenate.This method selects some key error-prone information bits for non-uniform segmentation check through Gaussian estimation.Effectively reduce the impact of error propagation,and at the same time concatenate with the cyclic redundancy check to select the correct path,which can improve the decoding performance of the decoding algorithm under the conditions of large lists and high SNR.Simulation results show that compared with CASCL decoding algorithms,the new parity-check concatenated code can improve the decoding performance by 0.1-0.15 d B on average under the same block error rate.In addition,the new concatenated code combined with the adaptive algorithms can take advantage of the performance improvement of the decoding algorithms to make the adaptive algorithms decode successfully in a smaller list,and reduce the decoding of the adaptive algorithms at a lower SNR by 6% to 25% complexity.In order to reduce the amount of computation in the decoding process of adaptive algorithms,this thesis proposed a fast adaptive algorithms based on path splitting set,which combines the search ability of path splitting and decoding algorithms to establish a selective splitting set.An important part of selective splitting is the selection of the split set.Combined with the characteristics of the adaptive algorithms,Fixed Split index Set(FSS)and Extended Split index Set(ESS)are designed.The FSS is selected by the channel,the ESS is selected by the path information,and an appropriate extended parameter is also selected.In addition,four types of special nodes are added for fast decoding.In special nodes,the search depth is determined by the number of split nodes in the calculation node.The simulation results show that this fast splitting algorithm combined with parity-check can further reduce the decoding complexity of the adaptive algorithm in low and medium SNR?...