RIS-Assisted DOA Estimation for Non-circular Signals Using a Single Antenna Receiver via Augmented Covariance Matrix Reconstruction

Reconfigurable intelligent surfaces (RISs) provide a cost-effective strategy to aid target sensing or direction-of-arrival (DOA) estimation in a non-line-of-sight (NLOS) scenario, where a RIS is employed to facilitate the receiver to sense the targets behind obstacles. The existing RIS-aided DOA estimation methods often overlook the pseudo-covariance statistical information of non-circular signals, leading to suboptimal estimation accuracy. To address this, we propose a novel RIS-assisted DOA estimation framework that synergistically exploits the asymmetric statistics and pseudo-covariance of non-circular signals. The challenge for single-antenna receivers to distinguish multi-source DOAs is tackled via a reflection coefficient adjustment mechanism termed RIS-ACMR (Reconfigurable Intelligent Surface-Augmented Covariance Matrix Reconstruction), which reconstructs the augmented covariance matrix from four solved submatrices. The reconstructed matrix enables direct application of standard noncircular DOA estimators. Simulation results illustrate that the proposed RIS-ACMR method achieves a significant performance improvement over the utilization of the standard covariance matrix for non-circular signals.