A Low Complexity LDPC Codes Decoding Algorithm

The typical scenario of the fifth generation of mobile communication system(5G)requires a quality of service with low delay and high reliability.Low-Density Parity-Check(LDPC)codes,in conjunction with soft-decision Sum-Product Algorithm(SPA),has been proved to approach the Shannon limit.However,SPA suffers from high computational complexity.To fulfill the low delay and high reliability requirement of 5G system,this thesis studies the low complexity hard-decision decoding of LDPC code.First,this thesis proposes a modified IMWBF decoding algorithm for binary LDPC codes.In IMWBF decoding under AWGN channel,the larger the magnitude,the higher the reliability.Therefore,for those abnormal bits(named outlier)having high receiving magnitude while wrong hard-decision,they will be very hard to be flipped during the iterations.To address this problem partly,this thesis proposes a modified IMWBF decoding algorithm through dynamic receiving magnitude update.In the proposed modification,the magnitude of the received signal is updated when the number of unsatisfied parity-check equations associated with this received signal is larger than a predetermined threshold.Through this modification,the outliers of the received signals can achieve a larger probability to be flipped than that in the conventional method.Simulation results show that the proposed modification has better performance than the conventional method with only a slight increase of the computational complexity.Second,it has been shown that nonbinary LDPC codes can offer better error-correcting performance than binary LDPC for short-to-medium codelength.Moreover,nonbinary LDPC is more suitable to combine with higher-order modulations.Therefore,this thesis studies the iteration joint detection-decoding(IJDD)algorithm.First,a modified IJDD algorithm is proposed to reduce the complexity.In the proposed modification,the information of the received signal is updated only when the number of unsatisfied parity-check equations associated with this received signal is larger than a predetermined threshold.The proposed algorithm can reduce the computational complexity without any loss of performance.Next,the multiuser LDPC coded system is studied and the iterative multistage IJDD algorithm with MMSE pre-processing is presented.Especially,for the Massive MIMO,the IJDD algorithm with low-complexity iterative sequential MMSE pre-processing is given.