Improvement Of Decoding Algorithm And Design For Raptor Codes

The reliability of information transmission has become a serious challenge to the communications business,with the increasing demand of big data and high rate communication.Channel coding technology is becoming more and more difficult in the communication environment.Fountain code is a new kind of channel encoding technology,compared the traditional encoding technology,the fountain code has the advantages: no rate constraint,low encoding and decoding complexity and no channel feedback.The initial fountain code is found on erasure channels,and a large number of research results show that fountain code has great performance in noisy channels,which has attracted extensive attention of researchers.In this paper,we will take the Raptor code in the fountain code as the research object,explore its encoding and decoding algorithm in the noisy channel,and design a kind of efficient Raptor codes.The main research contents are as follows:Firstly,the concept of digital fountain codes is introduced.The classical LT codes encoding process and decoding method in digital fountain codes are described in detail.Then,several degree distribution functions are described,and the performance of LT codes under different degree distribution functions is compared.Then,the coding principle of Raptor codes is introduced.The encoding and decoding methods of LDPC codes are analyzed,including the construction method of the check matrices,the encoding method and the two decoding algorithms,and the LLR-BP algorithm of Raptor codes is described.The performance comparison between Raptor codes and LT codes are emphasized.Secondly,the LLR-BP algorithm of Raptor codes is improved.In order to solve the problem of high complexity of the LLR-BP algorithm,The Min-Sum algorithm of Raptor codes is derived.The simulation results show that the Mim-Sum algorithm with correction factor has the approximately performance of the LLR-BP algorithm,and greatly reducing the amount of calculation,saving the hardware cost.In order to solve the problem that the iterative update algorithm is stepwise from the input nodes to the output nodes,a Raptor decoding method based on non dynamic scheduling is proposed.Compared with the traditional iterative method,this strategy can speed up the convergence rate and reduce the decoding delay.Then,combined with two improved methods,a non dynamic schedulingRaptor Mim-Sum algorithm is proposed,and its great performance is proved.Finally,an efficient Raptor code based on the optimization of the coding algorithm is proposed.In view of the deficiency of traditional encoding algorithm,an optimized encoding method is proposed.Compared with the traditional encoding method,the BER performance of optimized coding algorithm is significantly improved.A kind of high bit rate QC-LDPC codes is described,which is used as the precoding of Raptor code,then combined with the encoding optimization algorithm to construct an efficient Raptor codes.Compared with the traditional efficient concatenated codes,the great performance of the constructed Raptor codes is proved.