Subspace-Based Phase Self-Calibration for Robust DOA Estimation with MEMS Vector Hydrophones

Micro-electro-mechanical system (MEMS) vector hydrophones inherently suffer from inter-channel phase inconsistencies and structural non-orthogonality due to manufacturing imperfections, which significantly degrade the accuracy of Direction-of-Arrival (DOA) estimation. This paper proposes a low-complexity self-calibration method based on subspace normalization and second-order statistics, featuring a closed-form solution. The approach constructs a phase error model accounting for non-orthogonality, compensates for hardware-induced phase offsets via complex-domain de-rotation, and estimates the non-orthogonality angle by leveraging the statistical properties of ambient noise. By circumventing iterative optimization, the proposed method achieves unambiguous full-azimuth DOA estimation. Simulation and experimental results demonstrate that this technique effectively mitigates systematic DOA biases and delivers robust performance under moderate model mismatches. Furthermore, the limitations of the method under anisotropic noise conditions are analyzed.