针对宏小区稀疏码分多址技术（sparse code multiple access, SCMA）系统上行链路中远近用户到达基站存在的用户时延问题，提出了按不同时延进行用户分组的异步检测算法，利用前后符号间相同的信息进行迭代，形成了一种基于连续传播的消息传递算法（message passing algorithm based on serial propagation, SP-MPA）。另外，针对MIMO-SCMA系统中运用的部分边缘化检测算法造成的误码率性能下降的问题，还提出了一种基于动态阈值的部分边缘化消息传递算法（dynamic threshold partial marginalization algorithm, DTPM-MPA）。仿真结果表明，SP-MPA有效降低了异步传输的符号干扰带来的高误码率，并且DTPM-MPA能够在计算复杂度与误码率之间做出灵活平衡。

Aiming at the problem of delays of multiple users arriving at the base station in the uplink sparse code multiple access (SCMA) macro-cell system asynchronous detection algorithms were proposed with users grouped according to different delays and the same information between symbols was utilized and iterated, forming a message passing algorithm based on serial propagation (SP-MPA). Besides, aiming at the problem of worse bit error rate caused by partial marginalization algorithm used in MIMO-SCMA system, dynamic threshold partial marginalization algorithm (DTPM-MPA) was proposed. The simulation results show that the high bit error rate caused by symbol interference of asynchronous transmission is effectively reduced with SP-MPA and a flexible balance between computing complexity and bit error rate is formed with DTPM-MPA.