Robust Covariance Matrix Estimation for Uniform Rectangular Array
A. Ibrahim, C. Ren, Israël Hinostroza, Arnaud Breloy, Mohammed Nabil El Korso · 2026
Covariance matrix estimation is a fundamental component of adaptive signal processing methods. Motivated by the data structure brought by uniform rectangular arrays, we address the problem of block Toeplitz structure covariance estimation in non-Gaussian environments. We propose a robust estimator that incorporates structural constraints on the covariance matrix, enhancing resilience to deviations from Gaussian assumptions. The performance of the proposed approach is assessed with respect to the Cramér–Rao bound. A direction-of-arrival estimation for uniform rectangular arrays is presented to demonstrate the effectiveness of the framework in source localization tasks.