Research On Compressive Sensing Based Multiuser Detection Technology In Non-Orthogonal Multiple Access

Since the demand of the future 5G mobile communication is hugely increased in the aspects of spectrum efficiency and user connection quantity,the non-orthogonal multiple access(NOMA)technology breaks through the traditional orthogonal limit,which allows multiple users to share the same time-frequency resources via codedomain or power-domain multiplexing,becoming one of the key technologies of 5G.In order to save the signaling overhead and reduce the transmission latency,the uplink grant-free NOMA system has drawn much attention.However,the base station cannot obtain the user activity information before data transmission,and thus the user activity has to be detected.As the user activity in this system is sparse,satisfying the characteristic that the signal must be sparse in the compressive sensing theory.In view of this,this paper has done research on the compressive sensing based multi-user detection for uplink grant-free NOMA system,the specific contents are as follows:Firstly,considering that the base station is equipped with single antenna,we analyzed the impact of both the number of active users and the overloading factor on the performance of multi-user detection based on different types of compressive sensing algorithm in the uplink grant-free NOMA system.The results show that with the increase of the number of active users and the overload rate,the detection performance of the commonly used compressed sensing algorithm is getting worse,but the detection performance of structured iterative support detection(SISD)algorithm is still better.Secondly,considering that the base station is equipped with multiple antennas,we propose a compressive sensing based Hard Fusion Algorithm(CS-HFA)to achieve multiuser detection.More explicitly,we detect the user activity information by a conventional CS algorithm at each antenna,and then amalgamate the detected user activity information to derive an active user set.Finally,we can use the obtained active user set to estimate the active users' data.Compared with the traditional CS-based MUD in the NOMA system with single antenna at the BS,the proposed CS-HFA can achieve better BER performance by increasing the number of antennas appropriately.Thirdly,in view of the unknown number of active users in the uplink grant-free SIMO-NOMA system,we propose a sparsity adaptive matching pursuit based hard fusion algorithm(SAMP-HFA).Specifically,the algorithm firstly detects the user's activity information at each antenna by setting a suitable stop threshold for the traditional SAMP algorithm,and then amalgamate the detection information of multiple antennas to improve the accuracy of user activity information detection.The algorithm not only achieves multi-user detection with unknown number of active users in the SIMO-NOMA system with low complexity,but also increases the detection performance of the algorithm significantly as the number of antennas increases.