Font Size: a A A

Research On Non-Orthogonal Multiple Access

Posted on:2023-09-25Degree:MasterType:Thesis
Country:ChinaCandidate:B LiFull Text:PDF
GTID:2568306908450474Subject:Communication and Information System
Abstract/Summary:PDF Full Text Request
With the rapid development of Internet of Things,massive Machine Type Communications will be a typical communication scenario of the 5-th Generation Mobile Communication Technology(5G),and it has the characteristics of massive access but sporadic communication traffic.To meet the requirement of massive access and ultra-low latency,grant-free(GF)technology enables users to avoid the signaling scheduling process,which will cause high delay by a large number of signaling scheduling procedure,and nonorthogonal Multiple Access(NOMA)technology can use limited spectrum resources to provide services to more users,and higher spectrum efficiency compared with orthogonal multiple access.Therefore,grant-free non-orthogonal multiple access(GF-NOMA)will become one of the key technologies in the future communication systems.However,since the signaling scheduling process ignored,the base station requires to perform Active User Detection(AUD)first,and Data Detection(DD)after identifying active users.Based on the theory of compressed sensing(CS),utilizing the characteristics of massive access and sporadic communication in m MTC,the AUD is modeled as the sparse recovery problem,and the receiver algorithm of uplink GF-NOMA base station is designed for different scenarios in this thesis.The major contents are as follows:(1)For frame-wise sparsity communication scenario with perfect channel state information(CSI)at receiver,based on generalized approximate message passing(GAMP)algorithm,a joint AUD-DD algorithm for single antenna base station is proposed.The factor graph(FG)is designed by the idea of message passing(MP)to make full use of the frame-wise sparsity.Here GAMP algorithm,which used to solve single measurement vector based compressed sensing(SMV-CS)problem,is optimized by hybrid generalized approximate messaging framework with MP algorithm,to solve the multiple measurement vector based compressed sensing(MMV-CS)multiuser detection(MUD)problems.The simulation results show that the proposed Parallel GAMP-MP algorithm has significantly improved performance both in terms of AUD and DD compared with GAMP algorithm since MP algorithm is adopted to makes full use of frame-wise sparsity features and the statistical characteristics of transmission symbols.(2)For the communication scenarios without CSI,and dynamic block-wise sparsity with burst random access is considered,two joint AUD-CE-DD algorithms,Turbo-GAMP and SW-EM-GAMP are proposed,these algorithms utilize different strategies to optimize the performance of GAMP algorithm,where Turbo-GAMP is implement by extrinsic information in a turbo fashion,the performance of algorithm is optimized by turbo iterations between GAMP and MP.SW-EM-GAMP algorithm utilize block-sparsity to estimate user active probability by EM method.Moreover,the idea of sliding-window exhaustive strategy is adopted in MP algorithm to improve AUD accuracy,further improves the performance of channel estimation(CE)and DD.The simulation results show that the turbo-GAMP algorithm still has a good AUD performance even in scenarios that have more randomness of user active states.Based on the above research,the corresponding algorithms to solve the problem of GF-NOMA multi-user detection in both frame sparse communication scenarios with strong temporal correlation and dynamic sparse scenarios with relatively weak temporal correlation are proposed in this paper.The proposed algorithms achieve good performance with low complexity.
Keywords/Search Tags:grant-free non-orthogonal multiple access, multi-user detection, compressed sensing, generalized approximate message passing, message passing
PDF Full Text Request
Related items