Research On Denoising Signal Of MEMS Vector Hydrophone Based On Wavelet Adaptive

MEMS vector hydrophone has the advantages of high sensitivity,good directivity,small size and low cost.It has good anti-interference ability in complex underwater acoustic environment,and has achieved quite good results in positioning,orientation and signal processing.MEMS vector hydrophones have been successfully applied to underwater positioning navigation and underwater target detection.With the development of MEMS vector hydrophones,more and more applications will be applied in the field of underwater acoustic engineering,and the signal processing of MEMS vector hydrophones has become a research hot spot.Based on the systematic study of wavelet transform,adaptive algorithm and other related theories,this paper introduces wavelet transform theory into adaptive wavelet to form wavelet adaptive algorithm.The algorithm reduces the signal autocorrelation eigenvalue by orthogonal transformation of the input adaptive filter signal by wavelet transform,thereby improving the convergence and tracking speed of the adaptive filter.The characteristics and advantages of the wavelet adaptive algorithm are summarized from the point of view of the principle and application of the wavelet adaptive algorithm.Secondly,the adaptive algorithm LMS is discussed and analyzed,and the LMS is contradictory in terms of convergence speed and steady state.The LMS algorithm is improved by the variable step size method,and the convergence speed is fast and the steady state is good.Wavelet transform and variable step size LMS algorithm are used to improve the LMS algorithm,then wavelet adaptive algorithm is generated.The wavelet adaptive algorithm is applied to the MEMS vector hydrophone signal denoising process.The algorithm overcomes the disadvantages of the MEMS vectorhydrophone signal,such as the large dispersion of the eigenvalues of the adaptive filter input autocorrelation matrix,the slow convergence speed or the poor steady-state.Good results were obtained in simulation and real data experiment.