Combining ML and Compressive Sensing: Detection Schemes for Generalized Space Shift Keying

Abstract

Generalized space shift keying (GSSK) proposed for large scale MIMO systems promises significant advantages in terms of hardware costs. However, the maximum likelihood (ML) detection for GSSK is computationally intractable. Detection schemes using compressive sensing (CS) offer a favorable performance-complexity tradeoff. In this letter, we propose a novel GSSK detection framework, where CS algorithms are used to reduce the complexity of ML detection. The proposed framework significantly reduces the performance gap between existing CS-based detection schemes and ML detection while retaining the complexity advantage of CS algorithms. Simulation results demonstrate the superiority of proposed scheme over popular CS-based detection schemes.