Fast Estimation of mmWave Power-Angular Spectrum Using Leaky-Wave Antennas
Résumé
Leaky-wave antennas (LWAs) provide compact and cost-effective solutions for direction-of-arrival (DoA) estimation due to their frequency-dependent beam-scanning characteristics. However, in coherent multipath environments, the source covariance matrix becomes rank-deficient, which limits the performance of conventional subspace-based methods. Interpolationbased spatial smoothing (SSP) can restore the rank and recover DoAs, but accurate power estimation remains difficult, hindering the acquisition of the power angular spectrum (PAS). In this work, we propose an on-grid sparse Bayesian learning (SBL) approach that operates directly on the fast frequency-scanning responses of meandered waveguide-based LWAs. The method jointly estimates DoAs and their corresponding powers without requiring rank-restoration techniques. Simulation results demonstrate that the SBL-based approach provides robust and accurate PAS estimation, even in the presence of fully coherent sources.