Research On Enhanced Receiving Algorithm For Polar Coding System

Polar codes,as the first type of channel codes that are theoretically proven to achieve the capacity of any symmetric binary-input discrete memoryless channel,have attracted widespread attention.After more than ten years of development,polar codes have been adopted as one of the coding schemes for the enhanced mobile broadband communication scenario in the 5th generation wireless systems due to the well-developed coding constructions and decoding technologies.In this paper,based on the basic encoding and decoding algorithms of polar codes,two enhanced decoding methods are proposed.The bit-interleaved coded modulation with iterative decoding(BICM-ID)scheme for polar codes,along with the encoding and decoding schemes for polar-based turbo product codes(polar-TPCs)are also studied.First of all,the basic encoding and decoding algorithms of polar codes are introduced.We present the process of channel combining and channel splitting to introduce the concept of channel polarization and polar code construction.At the same time,several common methods to select the information bits are given.In the decoding algorithm part,the processes of some algorithms are introduced in detail,including the successive cancellation(SC)decoding,the successive cancellation list(SCL)decoding,the cyclic redundancy check aided SCL(CA-SCL)decoding,and the belief propagation(BP)decoding.The encoding and decoding processes of systematic polar codes are also introduced.By means of simulation,the performances of different encoding and decoding algorithms are analyzed.Simulation results show that the CA-SCL algorithm outperforms SC,BP and SCL decoding algorithms.Secondly,we propose two enhanced decoding methods,respectively based on the BP and CA-SCL decoding algorithms.The first one,Na-SBPL decoding algorithm,combines the noise-aided decoding method with the sparse graph based BP(Sparse-BP)algorithm,which employs the BP decoding algorithm of lowdensity parity check(LDPC)codes to decode polar codes by pruning the check matrix of polar codes to be sparse.The proposed Na-SBPL algorithm can effectively improve the performance of Sparse-BP algorithm,and makes it possible for polar codes and LDPC codes to share the same decoding architecture.A bit-flipping based decoding strategy called C-SCLFlip is also proposed,in which a limited number of unreliable bits are corrected by traversal to improve the overall decoding performance.Simulation results show that the CSCLFlip decoder outperforms the CA-SCL decoding with the same list size.It also shows competitiveness with some existing flipping-based SCL decodings while lower implementation complexity is needed.Then,a new BICM-ID scheme for polar codes is proposed,in which the C-SCLFlip decoder is employed.The C-SCLFlip decoder can provide a large number of hard decision outputs,from which extrinsic information can be calculated using Pyndiah's method.After being interleaved,the extrinsic information is sent to the demodulator as the prior information.Consequently,an iterative demodulation and decoding loop comes into being.In order to reduce the computational complexity,we analyze the reliability of the extrinsic information and find that only two paths in each SCL decoding attempt are needed to generate sufficient enough extrinsic information.Simulation results show that the proposed BICM-ID scheme can effectively improve the error correction performance.Compared with the BICM scheme employing CA-SCL decoding,a performance gain of 0.5 d B can be obtained.Meanwhile,the average calculation complexity is acceptable and is similar to a CA-SCL decoder at medium to high signal to noise ratio.Finally,the encoding and decoding schemes of polar-TPC are introduced.The component codes can be either systematic or non-systematic,and all row(column)component codes can be decoded in parallel.Row decoding and column decoding are performed alternately.A flipping-based C-SCLFlip decoder is employed to decode each constituent code and provide a larger code set,from which soft information can be obtained to update the decoder's input in the next iteration.Simulation results show that the performance of polar-TPC improves significantly with iterative decoding.After the same times of iteration,the proposed polar-TPC can outperform the BCH-TPC.The performance of polar-TPC can catch up with that of a polar code with the same code length and coding rate employing the CA-SCL(list size = 4)decoder,while the decoding delay is much lower.In addition,the proposed polar-TPC decoding scheme is compatible with any SCL decoding based acceleration algorithm,which can further reduce the decoding latency.