DOA Estimation of Multibeam Frequency Beam Scanning LWAs Based on Sparse Bayesian Learning

Abstract

Multibeam frequency beam scanning leaky wave antennas (FBS-LWAs) enable a compact system with reduced bandwidth but introduce parasitic interference that impairs the direction-of-arrival (DOA) estimation. To solve this issue, this letter proposes a novel DOA estimation method tailored for multibeam FBS-LWAs. Theoretical analysis reveals that the true DOAs align with multiple peaks in the radiation pattern. Based on this insight, a cropped hypercomplete set is constructed via grid refinement in the frequency-space domain. The sparse Bayesian learning (SBL) algorithm, enhanced by singular value decomposition (SVD-SBL), is then employed to suppress parasitic peaks and accurately recover the true DOAs. The simulation results demonstrate the proposed method’s effectiveness androbustness.