Research On Estimation Of Source Number And Adaptive Filtering Technology For Acoustic Vector Sensor Array

In most cases,estimation of source number is required to be implemented before direction of arrival(DOA)estimation,target tracking and target localization.In the presence of non-uniform noise,this thesis proposes a new MDL source number estimation and DOA estimation algorithm,which combines the MDL criteria with the genetic algorithm.The proposed method solves the problems that the traditional MDL algorithm and traditional MUSIC algorithm suffer from the performance degradation in the presence of the non-uniform noise.This thesis proposes a new variable step size for the LMS algorithm based on logarithmic function and sigmoid like function.In addition,the proposed method is compared with SVSLMS algorithm and the G-SVSLMS algorithm.The selection basis of constant step size LMS algorithm and variable step size LMS algorithm is analyzed as well.And then,used this variable step size for the LMS algorithm in adaptive noise canceller.In this thesis,for the small size array,we study the supergain technique to acquire high output gain of the acoustic vector sensor array.The correlation of isotropic noise for the acoustic vector sensor array is firstly analyzed.Then,the correlation matrix of noise is applied to calculate the weights of the supergain technique.Furthermore,this thesis applies the adaptive noise cancellation technology on the received signal of the acoustic vector sensor array in order to reduce the influence of noise.The performance of two adaptive noise cancellation methods for single vector hydrophone is analyzed.Moreover,the generalized sidelobe canceller is applied to the acoustic vector sensor array.Finally,the MVDR method is combined with the adaptive noise cancellation technology to improve the performance of DOA estimation.