Compressive Sensing-Based Low-Complexity Detector for GSSK-MIMO Systems

Abstract

The generalized space shift keying (GSSK) technique proposed for massive MIMO systems has significant advantages in terms of hardware costs. However, the maximum likelihood (ML) detection for GSSK is computationally intractable. Consequently many suboptimal schemes with varying performance complexity trade-offs like compressive sensing (CS) detectors are proposed. Nevertheless, the existing CS detectors estimate the transmitted signal only by single-stage, which results in a decrease in detection performance. In order to solve this issue, we proposed a novel detection framework in this letter. Different from the existing CS detectors, the framework includes two stages: rough and precise selection. In the rough selection stage, the Euclidean distance criterion is employed instead of the traditional inner product operation to determine an optimized superset of active antenna indices. The precise selection stage performs a further traversal search over the superset acquired by the first stage to obtain the final indices. The simulation results illustrate that the performance of the proposed scheme can exhibit significant performance enhancement over the existing CS detectors while retaining the low complexity advantage.