Error Recovery Based Low-Complexity Detection for Uplink Massive MIMO Systems

Abstract

In this letter, we introduce a new approach based on error recovery for detection in uplink massive multiple-input multiple-output (MIMO) systems. In the proposed work, first, the non-sparse massive MIMO system is converted into a quasi-sparse error (i.e., error vector consists of a small fraction of significant elements) system by using a rough initial estimate of the transmitted symbol vector. Next, the error is estimated by a low-complexity error recovery technique which detects the error elements in an ordered pattern. Finally, to obtain the output symbol vector, the estimated error vector is combined with the initial solution vector. Furthermore, in order to enhance performance of the algorithm, multiple iterations of error recovery technique are performed. Simulation results reveal that the proposed algorithm outperforms the recently reported methods with approximately equal complexity.