Fixed Point Number Decoding Implementation Of Polar Codes And Research On Its Application In BICM/SCMA Systems

Polar codes are channel coding schemes proposed by E.Arikan in 2009.As the code length increases,the error correction performance of Polar codes becomes better and better,even superior to the error correction performance of Turbo codes as well as Low Density Parity-Check(LDPC)codes.This paper is dedicated to Polar codes in the additive white gaussian noise(AWGN)channel,and it mainly includes the following five research points:1.The basic principle of channel polarization and polar coding are present in detail.This paper illustrates channel polarization effect.Besides,it introduces two construction methods for the AWGN channel.The first one is the Bhattacharyya parameter method.The second one is the Gaussian approximation method.Simulations show that Polar codes constructed by the second construction method perform better.2.The performance of Polar codes under the standard successive cancellation list(SCL)decoding algorithm is improved.This paper utilizes lazy-copy technique,pruning procedure and CRC precoding in order to improve the performance of SCL decoding algorithm.Meanwhile,it converts SCL recursive decoding algorithm into non-recursive decoding algorithm using stack.Simulations show that(1)SCL decoding algorithm with lazy-copy technique and pruning procedure can obviously reduce decoding complexity with very small performance loss.(2)SCL decoding algorithm with CRC precoding can gain more performance gain.(3)On the MATLAB,non-recursive decoding algorithm can gain more than 60% efficiency improvement than recursive decoding algorithm.3.The fixed-point number decoding schemes of SCL decoding algorithm are designed and realized.The first one is the quantization scheme based on automatic search algorithm.The second one is the quantization scheme based on approximate algorithm.This paper quantifies Initial log-likelihood ratio(LLR)value,internal LLR value and path-metric value respectively.Besides,it realizes the fixed-point number decoding in the BPSK and high-order modulation systems respectively.Simulations show that(1)the second quantization scheme has less quantization time and quantization bit numbers than the first one.(2)In the high-order modulation system based on the boundary decision method,as the modulation order increases,the optimal quantization bit numbers of internal LLR value and path metric decrease.4.A PCBICM system which introduces Polar codes into the bit-interleaved coded modulation(BICM)system is formed.This paper analyzes the random interleaving algorithm,the block interleaving algorithm and the LLR demodulation algorithm based on boundary decision method.Meanwhile,there is a comprehensive analysis about theperformance of PCBICM system with different modulation methods,code lengths and decoding algorithms.Simulations show that the decoding performance of the PCBICM system is far superior to that of the high-order modulation system.5.A PCSCMA system which introduces Polar codes into the sparse code multiple access(SCMA)system is formed.Through simulations of PCSCMA and Turbo_SCMA systems,we find: PCSCMA system perform better.Therefore,it makes sense to combine Polar codes and SCMA technology in terms of system performance.