Performance Analysis And User Pairing Of SCMA Based Hybrid D2D&Cellular Networks

The shortage of spectrum resources and the massive users access are two of the major challenges in next generation mobile communication systems(5G).Under such circumstances,Sparse Code Multiple Access(SCMA)was proposed as a new non-orthogonal multiple access scheme.Its non-orthogonality property allows users to overload,and the sparsity of its codewords ensures that the receiver can detect the user information with relatively low complexity.Therefore,SCMA is one of the promising multiple access candidates for 5G.On the other hand,D2D,as a direct communication technology which can effectively improve the system spectral efficiency,and is considered as one of the key technologies for 5G.This thesis aims at investigating the SCMA based hybrid D2D&cellular networks.First of all,in this thesis,the concept of spatial point process(especially HCP)is applied in SCMA based hybrid D2D&cellular networks,and the corresponding interference model is proposed.Based on the simulation platform built using Matlab,it is shown that the SCMA can achieve better performance compared with the OFDMA,in terms of the coverage probability,the average rate and the area spectral efficiency.Meanwhile,the simulation under different network topologies,including HCP,traditional hexagonal cellular structure and PPP,is also performed,showing that HCP is a more proper model that describes the tropology of the network.In addition,in this thesis,the user pairing algorithms of single-cell uplink scenario in the SCMA based hybrid D2D&cellular networks is studied.Based on the properties of SCMA codebook,an orthogonal user pairing algorithm is proposed to optimize the average sum rate of the system.The algorithm proposed has the advantage of low complexity when the codebook size is large.What is more,the simulation results show that the optimization algorithm proposed in this thesis can effectively reduce the interference and improve the average sum rate of the system,compared with the random user pairing algorithm and the greedy algorithm.