Grant-free non-orthogonal multiple access (GF-NOMA) based on multiple-input multiple-output (MIMO) has attracted much attention as a promising technique to support massive connectivity and bursty data transmission in massive machine-type communication. In this paper, we propose two compressed sensing based multiuser detection (MUD) algorithms for the MIMO-enabled GF-NOMA system. First, the spatially enhanced variational Bayesian inference (SE-VBI) algorithm is developed for MUD by exploiting the Gaussian mixture prior and diversity combining technique. Then, by applying the covariance-free (CoFe) strategy to the SE-VBI framework to estimate the diagonal elements of the posterior covariance, we propose a low-complexity MUD method named SE-CoFe-VBI. In particular, the proposed algorithms integrate the multivariate nature of the transmitted signal, i.e., discreteness, sparsity, and spatial correlation. Simulation results show that the proposed algorithms offer improved detection performance over the state-of-the-art spatially enhanced sparse Bayesian learning method.
Cost-Efficient VBI-Based Multiuser Detection for Uplink Grant-Free MIMO-NOMA
24.06.2024
1062760 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Uplink Performance Analysis of Grant-Free NOMA Networks
IEEE | 2022
|Multiuser Diversity Gain in Uplink NOMA
IEEE | 2018
|