| Low-density parity-check(LDPC)code is a class of channel coding scheme with excellent performance.For practices of LDPC codes,a key issue is the design of efficient LDPC decoders.In particular,when high-end chips are unvailable,it is a worthwhile way to break through the restrictions at the algorithm level and develop a low complexity and highperformance quantization decoding algorithm.Recently,the information bottleneck(IB)method,which originated from the field of machine learning and information theory,has been successfully applied to the design of quantization decoders for LDPC codes.The resulting high performance quantization decoder can approach the performance of floatingpoint SPA(Sum-Product Algorithm)decoder by using only 4 quantization bits.Moreover,the LDPC IB decoder deals with only unsigned integers and replaces complex check node operations by simple lookup tables,and thus is very suitable for practical implementations.This thesis mainly studies the optimal design and performance analysis of LDPC IB decoder.The main work and innovation are summarized as follows:(1)The basic principle and implementation algorithm of information bottleneck method are briefly reviewed,and the dedicated sequential IB algorithm for finding the optimum BPSKAWGN channel quantizer is introduced in detail.Then,based on these basic knowledge,the design method of IB quantization decoders for regular LDPC codes is introduced,mainly including the generation of check node IB lookup tables,the generation of variable node IB lookup tables and the selection of quantization decoder design signal-to-noise ratio.(2)Through the analysis,it is found that the number of table lookups is proportional to the square of the node degrees,which is unfavorable for LDPC codes with large node degrees,such as the finite geometry LDPC codes and high-rate LDPC codes.To deal with this issue,an improved scheme for the design of IB decoders is proposed based on the forwardbackward algorithm.The proposed scheme avoids duplicate message computations by careful use of the intermediate messages and thus makes the number of table lookups linear in the node degrees.Numerical results are provided to demonstrate the effectiveness of the proposed scheme.(3)The problem of message alignment(MA)in the design of IB quantization decoders for irregular LDPC codes is summarized.Message alignment problem will result in a notable performance degradation of LDPC IB decoder.Taking LDPC codes in the Wi MAX standard as examples,this thesis introduces two approaches to solve the problem of message alignment,namely MA algorithm from the perspective of information theory and the IB algorithm from the perspective of information bottleneck theory.Simulation results show that the IB algorithm has the same performance as MA algorithm in dealing with the message alignment problem of IB quantization decoder of irregular LDPC codes.(4)This thesis introduces a design method to solve the problem of punctured information bit of LDPC codes by IB method.Based on this design method,the IB quantization decoder of LDPC codes in 5G standard is designed,and the decoding performance of the IB quantization decoder of 5G LDPC codes is analyzed.The simulation results show that the design of 5G LDPC IB decoder is affected by punctured information bit and degree-one variable node,and the FER performance curve has a notable error floor. |
