Research On The Grant-free Meschanism Of Uplink SCMA System

In order to meet the needs of the fifth generation of mobile communication systems,sparse code multiple access(SCMA)as one of the key technologies have gradually been recognized and studied by people.SCMA has the ability of high overload and high throughput depending on its unique sparse codebooks.However,for the uplink SCMA system,the grant-free mechanism is still not consummate enough.This thesis aims to optimize the packet loss rate performance and decoding performance of the uplink grantfree SCMA system by studying the Contention Transmission Unit(CTU)allocation strategy,codebook allocation strategy and the Message Passing Algorithm(MPA).First of all,this thesis analyzes the research status of link-level and system-level,uplink and downlink SCMA systems in 5G at home and abroad,the basic model of uplink SCMA and the principle of grant-free.Then,the SCMA codebook design and the uplink grant-free multiuser detection technology are separately described and simulated.Then,the thesis describes the relationship among the three types of uplink factors: users,CTUs and codebooks.On this basis,the packet loss rate caused by the CTU conflict in the uplink SCMA system is deduced.With the minimum packet loss rate as the objective function,an iterative heuristic activity-based CTU allocation strategy is designed.The NP-hard problem is turned into a solvable problem.Simulation analysis shows that this method can greatly reduce the packet loss rate.Then,based on the result,the average progressive complexity of the codebook of the uplink SCMA system is deduced.With the minimum average progressive complexity as the goal,a codebook allocation strategy based on the activity degree is designed,which translates a nonlinear integer programming problem into a solvable quadratic 0-1 programming problem.Finally,we propose a serial decoding algorithm for the problem that the codebook conflict in the uplink grant-free system can bear but can not be avoided.Compared with the ideal condition,in the decoding algorithm,the codebook conflict situation will have a huge impact on the convergence of the MPA.So the external information transfer(EXIT)map is introducted to explore the convergence problem.Then based on the conclusions from external information transfer diagram,using the confidence coefficient of the codeword information in the decoding algorithm,the optimization design of the serial decoding algorithm under the codebook conflict is completed,and the decoding complexity is reduced without substantially affecting the decoding performance and convergence.