Design And Optimization Of Sparse Code Multiple Access Systems

5G,the fifth generation mobile communication,is one of the key technologies supporting the development of the future mobile communication system in China.In future,the 5G mobile networks will support the rapid growth of mobile data traffic with very limited spectrum resources,and Non-orthogonal multiple access(NOMA)is a key technology in 5G systems.It utilizes multiplexing technology to enable the same resource to be shared by more users,which greatly improving system capacity and spectrum efficiency.While Sparse Code Multiple Access(SCMA)stands out in those because of its codebook sparsity,which can meet the needs and developments of 5G and more cutting-edge technologies.Firstly,because the performance of the SCMA system can be improved by optimizing the designed codebook,so for the sender of SCMA system,this paper proposes an efficient codebook design based on one-dimensional search algorithm,aiming at minimizing the upper limit of pairwise error probability(PEP).Our design based on a new SCMA model,which expands the column space of its channel response matrix.And,by optimizing the settings for amplitude and phase rotation,each user will get the same minimum mutual information.Based on this premise,the upper limit of the PEP is optimized by a one-dimensional search algorithm.Furthermore,the performance of the proposed codebook is verified by the message passing algorithm(MPA)at the receiver.Compared with the existing codebook,the one-dimensional search codebook which minimizes the upper limit of pairwise error probability is better in performance and has better practical prospects.Secondly,for the multi-user scenario of SCMA system,the detection algorithm at the receiver end is improved and optimized.Based on the traditional successive interference cancellation algorithm(SIC),one study proposed a multi-feedback form that can overcome the error propagation phenomenon.By comparing the reliability of the local maximum likelihood and the preset threshold,the unreliable codeword is first screened out,and the high reliability codeword is preferentially decoded.For unreliable codewords,a set of candidate codewords is generated and the best codeword is selected therefrom.In this paper,on one hand,the selection of the candidate codeword set is adjusted.The original best path is filtered first,and then the competition path corresponding to the best codeword bit is reserved layer by layer.On the other hand,as the number of reserved paths increases,we take the likelihood ratio information of the candidate codewords as an input to the cascaded decoder,thereby improving the performance of the system output.