Detection of Spatially Modulated Signals via RLS: Theoretical Bounds and Applications

TL;DR

This paper analyzes the performance of regularized least-squares detection in massive multiuser spatial modulation MIMO systems, deriving theoretical bounds and proposing practical tuning methods validated by numerical experiments.

Contribution

It provides a theoretical characterization of RLS-based detection performance and introduces an efficient tuning approach for LASSO-type detectors in MIMO systems.

Findings

Derived asymptotic transmit rate and distortion bounds.

Quantified error rates for Bayesian optimal detection.

Validated theoretical results with numerical experiments.

Abstract

This paper characterizes the performance of massive multiuser spatial modulation MIMO systems, when a regularized form of the least-squares method is used for detection. For a generic distortion function and right unitarily invariant channel matrices, the per-antenna transmit rate and the asymptotic distortion achieved by this class of detectors is derived. Invoking an asymptotic characterization, we address two particular applications. Namely, we derive the error rate achieved by the computationally-intractable optimal Bayesian detector, and we propose an efficient approach to tune a LASSO-type detector. We further validate our derivations through various numerical experiments.