Fast Power Reconstruction for User Detection of Autonomous Grant-free Data-only Schemes

Autonomous grant-free data-only transmissions greatly simplify machine type communications, but present many challenges. One of them is the compressed sensing multiple measurement vector (MMV) problem of user detection. The sensing matrix is very underdetermined, and the amount of measurement vectors is large, which affect performance and complexity, respectively. In this paper, a fast power reconstruction (FPR) algorithm is proposed. FPR simplifies MMV to single measurement vector, and no iterations is required, which reduce the complexity. The row dimension of sensing matrix is expanded, and thus the performance is improved. The simulation results validate its good performance and ultra-low computational complexity.