Research And Implementation Of Microphone Array Sound Source Localization System

This investigates the impact of environmental noise and reverberation on the accuracy of sound source localization in audio conferences.To overcome the limitations of individual microphones in sound source localization,such as poor spatial resolution and weak interference resistance,microphone array systems are commonly employed for sound source localization and identification.The main focus of this study is to research the algorithms for sound source localization and develop a sound source localization system.The main research contributions of this paper are as follows:(1)Theoretical models for signal reception in sound source localization systems are studied.First,speech signals emitted by the sound source are collected and subjected to audio preprocessing.Then,the influence of microphone array structure,inter-element spacing,and the number of elements on the accuracy of the sound source localization system is investigated.Subsequently,near-field and far-field models,as well as ideal and practical models for speech signal reception by microphone arrays,narrowband and wideband signal models,and the reverberation model in indoor environments are introduced.(2)Time delay estimation algorithms in sound source localization based on time difference of arrival are studied.Generalized cross-correlation time delay estimation algorithms are investigated,and simulation analysis is conducted on four weighting functions:GCC,PHAT,SCOT,and ROTH.The characteristics of each weighting function are summarized.Based on these characteristics,an improved MLPT-ργ weighting function is proposed,and the optimal parameter values for the improved MLPT-ργ weighting function are determined through research.Finally,the improved MLPT-ργ weighting function is compared with SCOT,PHAT,and PHAT-ργ weighting functions through simulation analysis,demonstrating that the MLPT-ργ weighting function achieves the highest accuracy in time delay estimation.(3)Position estimation algorithms in sound source localization based on time difference of arrival are studied.Firstly,the target function-based spatial search localization method is investigated,and the original two-dimensional localization model is expanded to a three-dimensional model to enhance the practicality of the localization system.Secondly,the pentagonal array geometric localization algorithm is studied and compared with the tetrahedral array algorithm through simulation analysis.The results indicate that the pentagonal array structure achieves higher localization accuracy and eliminates localization blind spots.Finally,a BP neural network sound source localization model is established,and simulation analysis is performed on the pentagonal array model using MATLAB software.(4)A sound source localization system using a pentagonal array microphone is constructed.Firstly,the overall design scheme for the microphone sound source localization system is presented.Then,the hardware and software components of the sound source localization system are designed.Finally,experimental testing is conducted on the constructed sound source localization system,and the results of the position estimation are provided.Based on the experimental results,it is concluded that the constructed microphone sound source localization system meets the requirements for sound source localization in audio conferences in indoor environments.