Underwater DOA and SAP estimation method based on fast multi snapshot without inverse sparse bayes
By deriving the relaxed evidence lower bound in complex matrix form for underwater target detection, the high computational cost caused by matrix inversion in sparse Bayesian learning is solved, achieving high-precision and high-speed DOA and SAP estimation, which is suitable for online systems and a wider range of application scenarios.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HARBIN ENG UNIV
- Filing Date
- 2022-11-24
- Publication Date
- 2026-07-21
AI Technical Summary
In underwater target detection, existing technologies, such as sparse Bayesian learning (SBL), suffer from high computational costs due to matrix inversion, making them difficult to apply to large-scale problems and online systems. Furthermore, the applicability of single-snapshot and multi-snapshot beamforming models in the complex domain is limited.
We propose an underwater DOA and SAP estimation method based on fast multi-snapshot inverse-free sparse Bayesian. By extending the properties of the smoothing function to the complex matrix form, we derive the relaxed evidence lower bound (relax-ELBO), decouple the relationship between the target source and the measurement matrix, avoid matrix inversion operations, and use the variational distribution of latent variables to approximate the posterior distribution for Bayesian inference.
It reduces computational complexity, improves the accuracy and speed of DOA and SAP estimation, enables online applications, and extends to fields such as radar imaging, medical imaging, and seismic imaging.
Smart Images

Figure CN115859012B_ABST