Research And Implementation Of Sparse Code Multiple Access For 5G

Mobile internet and internet of things are important application scenarios for the fifth generation of mobile communication(5G)in the future.Rapidly developed mobile internet and internet of things demand mobile communication network to provide massive access service.Multiple access which is an important mobile air interface technology can improve subscribers efficiently.Therefore,researching and designing new multiple access has important research value and practical significance.Sparse code multiple access(SCMA)which is a new non-orthogonal multiple access can be a good fit for application scenarios of 5G.Thus it becomes a concern in industry and academia.Before standard of 5G is settled,this thesis researches multiple access for 5G based on SCMA,aiming to design appropriate multiple access scheme for 5G.Work of this thesis focuses on design criterion for codebook and low complexity decoding.Firstly,the thesis analyses the method of designing codebook based on shuffling multidimensional constellations in real and imaginary axes.According to the basic idea of optimal design for codebook,the thesis proposes a new design criterion for codebook.Simulation shows that codebook with new design criterion has great advantages than codebook using the design criterion that is desinged by shuffling multidimensional constellations in real and imaginary axes in system BER.As optimal method of designing codebook is still unkown,the new design criterion for codebook is a good alternative scheme for SCMA.Then,the thesis proposes logarithmic domain factor subgraph-based message passing algorithm(Log SG-Based MPA)after researching original message passing algorithm(MPA)based on normal factor graph.On the one hand,Subgraph with symmetrical structures is instead of the original factor graph for message passing algorithm,which reduces effective operation of function nodes in the process of message iterations.On the other hand,message passing algorithm is introduced to the logarithmic domain,which eliminates complexity operation by Jacobian logarithm.Analysis and simulation show that Log SG-Based MPA has great advantage than MPA in low-complexity decoding which is at the expense of small system BER performance.Therefore,Log SG-Based MPA can have a good balance between computation complexity and BER performance.Finally,this thesis proposes a FPGA implement scheme based on the modified decoding algorithm for SCMA decoder.Verilog is used to design a low-complexity SCMA decoder circuit with six users.The result of synthesis shows that the consumption rate of logic resource is about 17%.The test result shows that SCMA decoder needs 246 clocks in one decoding process.And test BER is in agreement with simulated BER,which meets design requirements.