Research On Rate Compatibility And Decoding Method Of Polar Code

Under the guidance of Shannon's theorem,channel coding technology is changing with each passing day.Turbo code and LDPC(Low density parity check,LDPC)code,two outstanding coding technologies,have not reached Shannon's capacity.The polar code proposed by Arikan is theoretically proven to be in B-DMC(Binary Discrete Memoryless Channel,B-DMC)can "reach"Shannon's capacity,and is adopted as a channel coding technical solution by 5G communication technology.Although polar code have lower encoding and decoding complexity and better error performance,they also have many defects.In terms of encoding,the length of polar code is limited to an integer power of two.In terms of decoding,The decoding performance of short and medium code lengths is not good enough.Therefore,aiming at these two issues,this article studies the rate compatibility of polar code and the improvement of decoding methods.The main innovations are as follows:(1)Aiming at the problem that the code length of the polar code is limited,a shortening scheme of system polar code based on the assistance of Gaussian approximation is proposed.This scheme aims to preserve the polarization effect,uses Gaussian approximation to assist in screening poor sub-channels as shortened channels,and obtains a more excellent rate-compatible polar code through systemization.The simulation results show that in the AWGN(Additive White Gaussian Noise,AWGN)channel,the error performance of the proposed shortening method is better than both the shortened polar code and the shortened polar code based on Gaussian approximation-aided,and it is also better than the QUP(Quasi Uniform Puncturing,QUP)of the system polar code.When the BER(Bit Error Rate,BER)is 10-4,the proposed shortening method obtains a gain of about 0.3 5dB than the shortened polar code based on Gaussian approximation-aided,and a gain of about 0.45 dB than the shortened polar code.(2)For polar code with medium and short code lengths,some sub-channels have incomplete polarization,which increases the possibility of information bit errors during transmission.In order to solve this problem,a BA-SCL(BCH-aided Successive Cancellation List,BA-SCL)decoding method is proposed.This method cascades the BCH(Bose-Chaudhuri-Hocquenghen,BCH)code and polar code,uses the segmented check method to correct the error bits,and combines the decoding method with the system polar code to further improve the error performance of the limited code length.The simulation results show that the performance of the decoding algorithm proposed in this paper is better than that of the BC-SCL(BCH-CRC Successive Cancellation List,BC-SCL)decoding algorithm.And the combination of the algorithm and the system polar code can achieve better error performance.When the bit error rate is 10-3 and the code length is 64,the proposed decoding algorithm based on BCH concatenated system polar code obtains about 0.2dB gain compared with he proposed decoding algorithm based on BCH concatenated polar code,and about 0.45dB than BC-SCL.