Sparse Bayesian learning underwater sound source localization method based on matched beams

The invention discloses a sparse Bayesian learning underwater sound source localization method based on matched beams, which aims at the problem of environment mismatch existing in the existing sparse Bayesian learning matching field localization algorithm, and utilizes a beam forming technology to convert sound pressure data received by an array into a beam domain space. Acoustic signals in a specific direction are limited or tracked in a beam domain, a positioning problem in the beam domain is converted into an underdetermined equation solving problem with sparse constraints, and finally iterative solution is performed through a sparse Bayesian learning updating formula. Compared with an original method, the method has the advantages that (1) the high-resolution low-sidelobe positioning result is ensured, meanwhile, higher tolerance to environment mismatch is achieved, and the positioning robustness is effectively improved; and (2) the operation speed is higher. 本发明公开了一种基于匹配波束的稀疏贝叶斯学习水下声源定位方法，针对现有的稀疏贝叶斯学习匹配