Research On Acoustic Source Location Algorithm Based On Microphone Array

The basic principle of the sound source localization algorithm based on the microphone array is to use a certain number of microphones to build a geometric array,used to collect and process the signal of the sound source,and ultimately achieve the target position determination.However,in practical applications,due to the presence of environmental noise and reverberation interference,the positioning results of the sound source localization algorithm have a certain degree of deviation.Therefore,in order to solve the problems existing in the actual situation,this paper focuses on the analysis of several traditional positioning algorithms,and based on this,proposes an improved algorithm that is suitable for practical applications.The theoretical model of the microphone array is studied,including its spatial model and signal receiving model.The room impulse response is calculated using the point source in the IMAGE method.This paper mainly focuses on the MUSIC algorithm and the time delay estimation algorithm.The traditional MUSIC algorithm is suitable for far-field narrow-band signals and has poor signal positioning performance for near-field wideband signals.This shortcoming was improved to achieve positioning of near-field three-dimensional sound sources.The time delay estimation algorithm consists of two parts: delay and positioning.In the time delay section,the generalized cross-correlation method and LMS adaptive method are mainly analyzed.The Matlab simulation analysis is used to compare the performance of the two methods.Aimed at the shortcomings of the traditional GCC algorithm,an improved generalized CSP algorithm is proposed.The research shows that the algorithm has lower computational complexity,higher anti-noise performance,and stronger anti-reverberation performance.In the localization section,the maximum likelihood method,angle distance location method,spherical interpolation method and linear interpolation method were analyzed.Simulations were performed on the simulation platform of LabVIEW.The results showed that the spherical interpolation method has a higher value when the number of microphones is enough accuracy.In this paper,the LabVIEW simulation platform is combined with the Simai Kehua USB-4222 data acquisition card to compile an audio sampling program to realize a sound source localization system that can collect and locate in real time under real environment.Experiments show that the improved generalized CSP method combined with the spherical interpolation method can obtain more accurate positioning results.