Research On Grant-free Non-orthogonal Multiple Access Technology Based On Subspace Pursuiting Algorithm

With the rapid development of the Internet of Things,the fifth-generation network must support huge connections between devices and consumers.The non-orthogonal multiple access(NOMA)solution can meet the needs of 5G giant connections,and therefore becomes the multiple access solution for future 5G networks.Grant-free NOMA has lower signaling overhead than scheduling-based NOMA,and is particularly suitable for short data transmission in the Internet of Things.However,Grant-free NOMA requires simultaneous user activity detection and data detection at the base station(BS)side,which makes multi-user detection a key issue for Grant-free NOMA.This paper studies the effective multi-user detection of NOMA without scheduling based on the subspace pursuiting algorithm.This paper first proposes a low-complexity parallel adaptive subspace pursuiting algorithm,which converts the multi-measurement vector problem into multiple single-measurement vector problems through parallel processing,and gives the corresponding information exchange strategy to make full use of Block sparsity.The simulation results show that,compared with the existing block sparse adaptive subspace pursuiting(BSASP)algorithm,the proposed parallel adaptive subspace pursuiting algorithm can achieve a great reduction in computational complexity at the expense of less detection performance.Next,the idea of multipath matching pursuiting algorithm based on depth-first search is applied to parallel subspace pursuiting algorithm,combined with the concept of regularization,a regularized parallel multipath subspace pursuiting algorithm is proposed.The simulation results show that the regularized parallel multipath adaptive subspace pursuiting algorithm can further approximate the performance of the BSASP algorithm.