Sparse Code Multiple Access

In mobile communication systems, multiple access technique is used to separate different users. However, considering future radio access in the 2020s, furture enhancement to achieve significant gains in capacity and system throughput performance is a high priority requirement in view of recent rapid development of mobile internet and internent of things (IoT).As a representative of non-orthogonal multiple access, the sparse code multiple access (SCMA) technique has many advantages. Firstly, overloading increases overall throughput and connectivity. Secondly, sparsity helps to reduce the complexity of detection. Finally, multi-demensional codewords bring shaping gain and better spectral efficiency. All in all, SCMA is a promising multiple access technique in the fifth generation mobile communication systems (5G).In this thesis, we proved that the update rule of variable nodes in the message passing algorithm (MPA) based SCMA receiver is based on Dempster's combination rule. We pointed out that the MPA based receiver is easily affected by the rusults against the commense (RACS) and simulations verified our analysis. In order to avoied the RACS, we can simply set a lower bound to the chip-messages. This algorithm is efficient in stopping the MPA based receiver from becoming divergent. However, the threshold can also stop the correct chip-messages, which deteriorates the performance of SCMA receiver. Furthermore, we propose a new "turbo-scheduling" algorithm, namely, the row message passing (RMP) algorithm, to reduce the occurrence of the RACS. In the RMP algorithm, chip-messages updated during the current detecting iteration are immediately utilized within the same iteration for calculation of new variable-messages. In other words, the turbo-effect is applied to the update of variable nodes. Thus, the reliability of the variable-messages improves and the convergence rate becomes faster.SCMA is a multi-user detection technique, and it can be jiontly designed with decoding techinique and multiple-antenna techinique. In this thesis, we studied the joint design problem, including the cascade SCMA-LDPC structure, the turbo SCMA-LDPC structure and the joint sparse graph (JSG) structure. When there are several uncorrelated receive antennas, we propose a new combination algorithm, named factor graph combining (FGC) algorithm. The FGC algotithm can obtain the gain of equal gain combining (EGC) algorithm, also the FGC algotithm can achieve additional gain of the connection degree of the variable nodes. Hence, the FGC algotithm is an alternative combining algorithm of EGC algorithm.Finally, we summary the research methods and achievements of this thesis and list some possible improvement ideas.