Research On Acoustic Vector Sensor Array DOA Estimation Algorithm Based On High Order Cumulants

Acoustic Vector Sensor is a new type of voice signal detection equipment which came into being recent years.The acoustic vector sensor can synchronously detect the sound pressure information and the three-dimensional velocity information of a point simultaneously,which will carry more useful information and detect more sensitive and wider range comparing with traditional sound pressure sensor.It is taking the place of traditional sound pressure sensor step by step in the domain of communication,radar,orientation,Biomedical Science,military defense and so on.Therefore,both from the national economy,or national security field,the research on acoustic vector sensor has a very important significance.The research on acoustic vector sensor technology is mainly separated into two major parts: one part focus on the mechanical structure of Acoustic Vector Sensor and another focus on the application of the sensor,namely,the received signal processing of the sensor.If the former study is the construction of the skeleton,the latter study is to give this skeleton soul.Direction Of Arrival,as a important subject of the array signal processing,is also a significant field of Acoustic Vector Sensor.It is the foundation of subsequent application to accurately estimate the DOA.Since the advent of Acoustic Vector Sensor,many algorithms based on second order statistics were proposed,such as A-MUSIC,A-ESPRIT and etc.,but still face a fatal problem: Gaussian colored noise suppression.Since the actual conditions are Gaussian colored noise,this problem seriously restricts the Acoustic Vector technology transformed into practical application.High-order cumulants,due to its insensitivity to Gaussian noise,once put forward was quite popular among experts and scholars.Many related algorithms were put forward and had achieved some good results,even some of them were put into application.But due to the Due to the huge amount of computation,The high-ordercumulants algorithm associated with the acoustic vector has never been mentioned.For the above reasons,in this paper,the DOA estimation of acoustic vector arrays based on high order cumulants is studied.First,from the point of view of noise suppression,the high-order cumulants DOA estimation algorithm of acoustic vector array is deduced.Secondly,the acoustic vector MUSIC algorithm based on higher order cumulants diagonal slices is proposed considering the problem of computational complexity in practical application.Furthermore,in order to obtain more accurate DOA estimation effect,an acoustic vector MUSIC algorithm based on high order cumulants 3 and 4 dimension slice is proposed.Then,in order to further reduce the computational complexity,the ESPRIT algorithm of high order cumulants diagonal slices of acoustic vector array is proposed without introducing ESPRIT algorithm into high order cumulus slices.Then,in order to overcome the problem that the DOA estimation accuracy is reduced in the ESPRIT calculation of high-order cumulants diagonal slices,this paper proposes an improved ESPRIT algorithm for high-order cumulants diagonal slices of acoustic vector arrays.Finally,the algorithm is verified by MATLAB,and the performance of DOA estimation is obtained.In this paper,we focus on the DOA estimation of acoustic vector sensor arrays,to expect that the research in this paper can be helpful to the theoretical research and practical application of acoustic vector array signals.