Introduction Underwater acoustic direction-of-arrival (DOA) estimation is fundamental to submarine detection, offshore exploration, and autonomous underwater vehicle navigation, where accurate source localization under severe multipath propagation and low signal-to-noise ratios remains challenging. Existing methods face critical limitations: traditional uniform arrays require extensive physical sensors to achieve sufficient spatial resolution, resulting in high hardware costs; vector sensor arrays typically neglect sparse array geometries that enable virtual aperture expansion; and conventional matrix-based processing discards multidimensional structural information through vectorization, leading to suboptimal performance. Methods To address these challenges, this paper presents a novel tensor-based DOA estimation framework that integrates hybrid scalar-vector sensor arrays (HSVSAs), fourth-order tensor modeling, and higher-order singular value decomposition (HOSVD). The HSVSA architecture combines vector and scalar sensors in a hybrid configuration to create virtual sensors, improving degrees of freedom while reducing hardware complexity. The tensor model preserves spatial–polarization coupling, enabling robust subspace estimation without iterative optimization. Results Simulations demonstrate that the proposed method outperforms conventional approaches. Discussion The proposed framework offers a practical solution for resource-constrained underwater acoustic applications, with potential for further optimization in real-world scenarios.
Zhang et al. (Fri,) studied this question.