Research Of Sound Source Localization Algorithm Based On Microphone Array

With the development of array signal processing technology,speech signal processing based on microphone array has become a new research hotspot.The microphone array adds a spatial domain based on the time domain and the frequency domain,which not only has the function of beamforming,but also can locate the sound source.Because the speech signal is a non-stationary wideband signal,coupled with the existence of reverberation,external background noise and other factors,the traditional sound source localization algorithm has the disadvantages of large computational complexity and low positioning accuracy.Based on the in-depth study of the sound source localization technology of the microphone array,this paper proposes a wide-band focusing 2D-IMUSIC algorithm based on GSA algorithm to better solve the problems of low positioning accuracy and high computational complexity.The main contents of this article are as follows:1)The preprocessing of speech signal,including pre-filtering,windowing framing,frequency division processing,speech endpoint detection,etc.,is analyzed.The model of array signal processing is introduced,which lays a foundation for the theoretical analysis of subsequent chapters.2)The spatial model of microphone array with different structures is researched.Some existing spatial spectrum estimation algorithms are analyzed and simulated.The simulation results show that the MUSIC algorithm has the dual advantages of computational complexity and estimation accuracy.3)For the traditional 2D-MUSIC algorithm with poor performance under low SNR,a 2D-MUSIC improved algorithm based on uniform circular array(2D-IMUSIC algorithm)is presented,and 2D-IMUSIC algorithm is used with 2D-MUSIC algorithm is simulated and compared.The results show that the 2D-IMUSIC algorithm can also have good estimation performance at low SNR.Then,for the shortcomings of 2D-IMUSIC algorithm unable to process coherent signals,this paper introduces a wideband focusing algorithm and gives a wideband focusing 2D-IMUSIC algorithm.Finally,in view of the large computational complexity of the wideband focusing 2D-IMUSIC algorithm,this paper considers the method of combining the broadband focusing 2D-IMUSIC algorithm with the GSA algorithm,and proposes a wideband focusing 2D-IMUSIC algorithm based on the GSA algorithm.The simulation results show that The proposed algorithm reduces the amount of computation under the premise of ensuring the positioning accuracy.4)Hardware and software design of the sound source localization system.Based on TMS320C6748 DSP to build a sound source positioning system test prototype,the system's processing module and voice acquisition module were designed.Finally,the system is tested by Microsemi test software.The test results show that the positioning effect of the sound source localization system can meet the expected requirements.