Research On The Low-complexity Multiple User Detection Of SCMA System

Sparse code multiple access (SCMA), is one of the strong candidates for the multiple access techniques in the 5th generation mobile communications. Evolving from low-density-signature multiple access, SCMA uses the codebook, i.e. the combination of multiple-dimensional constellations and low-density signature, to achieve overloading, making the number of users more than the quantity of the resources. Good multiple-dimensional constellation design ensures lower near-far effects and detection complexity, while massive conflicts are avoided by the usage of low-density signature, leading to excellent performance.Firstly, the architecture of SCMA system and the principle of the transmitter and receiver are introduced in this thesis. They take advantages than the counterparts in LDS-MA system. Then, the critical techniques in physical layer are studied, such as the codebook design at the transmitter side, especially several ways of multiple-dimensional constellation design. Existing techniques like maximum likelihood (ML) and message passing Algorithm (MPA) multiple-user detection, which are applied in the receiver of SCMA system, face with the challenge of increasing complexity when more users accessing the system.The methods of multiple user detection in SCMA system are similar to the ones in MIMO system. After the research on the detection algorithms in MIMO system, using subspace is proved to be a strong method to reduce complexity. The simulation results show these detection algorithms can keep balance between complexity and performance.Inspired by the subspace marginalization with interference suppression (SUMIS) detection, a novel low fixed complexity multiple-user detection technique for SCMA system is studied in the thesis as SUMIS-MPA multiple user detection. The simulation results shows that SUMIS-MPA detection reduces the search space, and keeps balance between the complexity and performance.