A modified near-field target localization method based on vector diagonal loading

Addressing the issue of poor positioning performance of the near-field MVDR (Minimum Variance Distance Reciprocal) algorithm due to limited sampling data, this study proposes a combination of vector hydrophones and an improved angle loading technique to enhance the robustness of the algorithm. The reasons for the degradation of positioning performance of the MVDR algorithm with a small number of snapshots are analyzed from two perspectives: optimal weight vector and output power spectrum. Firstly, the traditional angle loading algorithm is utilized to reduce the degree of noise feature diffusion, and then the loaded feature values are squared to obtain an improved covariance matrix, effectively resolving the issue of resolution degradation associated with traditional angle loading methods. Finally, the effectiveness and stability of the algorithm are verified through simulations and experimental data. The results demonstrate that the improved angle loading algorithm proposed in this paper performs comparably to traditional methods but significantly enhances resolution, making it more suitable for the positioning of underwater acoustic near-field targets.