Research On Indoor Sound Source Localization Method Based On Microphone Array

Because the disaster site may be filled with a large amount of dust,or the visibility of the site is low due to climate,power failure and other reasons,the acoustic wave is taken as the research object,and an indoor sound source location technology is proposed.The noise and reverberation of the environment play a key role in the positioning accuracy.By investigating the development of microphone array positioning technology,considering the computational complexity,positioning accuracy,indoor environment and other factors,the method based on time of arrival difference is adopted as the main algorithm in this thesis.The main research contents are as follows:(1)Based on the indoor IMAGE reverberation model,a weighted generalized cross-correlation delay estimation algorithm with improved wavelet threshold denoising is proposed.While preserving the characteristics of the common wavelet threshold function itself,an improved threshold function is proposed so that it can modify the threshold according to the distribution of noise,preserve the local characteristics of the signal and have a certain continuity.The generalized quadratic cross-correlation delay algorithm is used to improve the weighted function of PHAse Transform(PHAT),and the weighted function factor is introduced to detect the peak value of the signal,and the time delay between two signals is obtained.Experiments show that the algorithm has good anti-reverberation and anti-noise effects.(2)Aiming at the problems of low positioning accuracy and large amount of computation of the current positioning algorithm,a time-delay based positioning algorithm named controlled Response Power with PHAse Transform(SRP-PHAT)is proposed.In this algorithm,the eight-element circular microphone array is divided into four groups to form two groups of square microphone array.The search area is divided into eight regions,and the time-delay relationship between the eight regions is derived.Then the delay value of the microphone pair is calculated by the improved algorithm,and the real sound source area is determined.The azimuth Angle is found by SRP-PHAT algorithm.Experiments show that the algorithm can reduce the search space,reduce the amount of computation,and avoid the interference of false peaks caused by reverberation to a certain extent.(3)Based on theoretical simulation,the sound source positioning device is built and related experiments are carried out.The experimental results show that the combination of the delay estimation algorithm based on wavelet threshold denoising and the SRP-PHAT algorithm based on improved search area has higher accuracy,stronger anti-reverberation ability,better anti-noise performance and shorter running time than the target function space search algorithm based on time delay,which can meet the requirements of real-time location in disaster relief environment.