Research On Optimization Of Joint Iterative Detection And Decoding Algorithm For Polar-SCMA System

As a control channel coding scheme under 5G standard,polar code has an excellent error correction performance,moreover,compare with Turbo code and LDPC(Low Density Parity Check)code,it has lower encoding and decoding complexity.Facing the next generation of mobile communication demand for massive connection,sparse code multiple access(SCMA)technology directly maps binary coding bits to high dimensional code of the complex domain,and transmits the codes from different users in the form of sparse spread of nonorthogonal overlap to serve more users,while maintaining the same resource range.A SCMA system with polar code as the error control code can significantly improve the performance of the communication system,moreover,an ideal receiver design is very important to improve the whole quality of communication.In this paper,we study the receiver detection and decoding algorithm of polar coded SCMA joint system,and propose three optimal Polar-SCMA joint detection and decoding algorithms.Firstly,this paper introduces the background technology and theory technology of polar code and SCMA,respectively.The channel polarization and encoding process of polar code are studied.Several classical decoding algorithms of polar code,such as SC,SCL,BP and SCAN,are introduced in detail,and their performance is simulated and compared.The optimal algorithm is selected as the decoding method of the Polar-SCMA receiver.In addition,we give the encoding method of SCMA and It's classical MPA detection algorithm.Secondly,we propose a joint iterative detection and decoding algorithm S-JIDD for Polar-SCMA system based on the serial strategy,which merges the update process of user nodes into the update process of resource nodes.The update process of user nodes is no longer included in the message update process,so that the message is only transmitted between the resource nodes and the SCAN factor graph.Simulation results show that the proposed S-JIDD algorithm can reduce the computational complexity by about 16%?25% through accelerating the convergence speed,while maintaining an ideal BER performance.Furthermore,based on the S-JIDD algorithm,by studying the message transmission and update process of the Polar-SCMA joint factor graph,we propose a dynamic factor graph update algorithm PI-JDD based on the pruning iteration.Due to the different convergence rates of each branch in the factor graph during the iteration process,the confidence stability of each branch was judged before each iteration,and the branches that had converged were removed from the factor graph in time,so that they would not be calculated in the subsequent iterations.The complexity can be reduced at least 24% by pruning the factor graph.To further reduce the BER,this paper introduces an early termination mechanism at the end of each iteration,carries out CRC check on the decoded words at the end of each round of joint iteration and outputs the correctly decoded words timely,which can avoid the decoding error caused by decoding deviation during iteration.However,if the code does not pass the CRC check,the iteration will continue.The simulation results demonstrate that the proposed C-JIDD algorithm can significantly reduce the BER,and the error performance of 3 iterations of C-JIDD outperforms that of the original algorithm 5 iterations.Finally,the proposed algorithms are combined in this paper,and the simulation comparison is carried out under different channel conditions.The results demonstrate that the joint optimization algorithms can effectively reduce the computational complexity while improving the system error performance,and achieve a good balance between the complexity and BER.