Design of near-optimal local likelihood search-based detection algorithm for coded large-scale MU-MIMO system

Massive multiuser multiple input multiple output (MU-MIMO) system is aimed to improve throughput and spectral efficiency through a large number of antennas incorporated at the transmitter and/or receiver. However, the MU-MIMO system usually suffers from interantenna interference (IAI) and multiuser interference (MUI). The IAI imposes due to closely spaced antennas at each user equipment (UE), and MUI is enforced when one user comes under the vicinity of another user in the same cellular network. Most of the previous literatures considered any one of these interferences. However, the present work proposes singular value decomposition (SVD) precoding-assisted user-level local likelihood ascent search (LLAS) algorithm to mitigate both IAI and MUI. In the uplink MU-MIMO, the IAI is cancelled by SVD, and the residual MUI is mitigated by LLAS detection. The LLAS detection balances the trade-off between the classical suboptimal likelihood ascent search (LAS) and optimal maximum likelihood (ML) detection techniques. The proposed LLAS performs local search among all 2M(T)-dimensional neighborhood vectors at each UE, where M-T represents number of transmitting antennas of each UE. Thus, its performance is near optimal, and its complexity is much lower than ML detector.