Comparison of DOA Estimation Methods Based on Sparse and Noncircular Covariance in a Challenging Environment

Abstract

This paper evaluates and compares the performance of the MUSIC (Multiple Signal Classification), CMSR (Covariance Matrix Sparse Representation), and NC-CMSR (Non-Circular Covariance Matrix Sparse Representation) methods under complex conditions. We analyze their behavior in scenarios characterized by a limited number of sensors, highly noisy environments, and with a small number of snapshots. The study focuses, particularly, on the evolution of the Root Mean Square Error (RMSE) for direction-of-arrival (DOA) estimation as a function of the signal-to-noise ratio (SNR) and the number of snapshots, as well as on the detection probability with respect to SNR. The results highlight the relative robustness of the CMSR and NC-CMSR approaches under these constrained conditions, while also emphasizing the benefit of tailored processing for non-circular signals.