Research Of Sound Source Localiotn System Based On FPGA

The sound source localization technology based on the microphone array uses the microphone array to collect the signal of the target sound source,calculating the information of sound source by combining with the sound source localization algorithm after processing the signal.With the rapid development of digital array signal process,the localization of sound source signals through microphone array has been widely used,such as gun sound localization,individual military operations and other military fields,as well as smart speakers,hearing aids and other civilian fields.This thesis mainly focuses on the sound source localization system based on FPGA,with the microphone array as the signal acquisition front-end to collect signals and FPGA as the control core,realizing the sound source localization algorithm by combining with FPGA's high-speed parallel computing,flexible configuration and rich IO resources.This can be used to identify the direction of sound source and can also be applied in many other fields under the expansion of technology at a later stage.The work of this thesis focuses on the sound source localization system,including research on microphone arrays,simulation comparison of sound source localization algorithms,and design of FPGA-based sound source localization system.As a result,a sound source localization system is built on FPGA and experimental verification is carried out.Firstly,different types of microphones are compared and studied in many ways,and then the spatial models of linear microphone array and uniform circular array are theoretically deduced and analyzed,confirming uniform circular array as the signal collecting front-end of sound source localization system.The hardware peripherals of the six-microphone uniform circular array based on the MEMS microphone are designed,including power conversion circuit,microphone circuit and the interface circuit of FPGA,which has realized the synchronous acquisition of the six-channel sound signals.Secondly,three sound source localization algorithms of TDOA,MVDR and MUSIC are simulated on MATLAB,and the performance is compared and analyzed.The MUSIC algorithm with high resolution and excellent estimation performance is selected as the sound source localization algorithm realized on FPGA.Then,the design of the entire system is completed on FPGA.Sound source localization system based on FPGA contains microphone array signal collection and processing and the MUSIC algorithm,based on microphone array hardware,the signal acquisition,data bits reduction,filtering,data frame are completed,and each part of the MUSIC algorithm such as calculation of correlation matrix,the matrix eigenvalue decomposition,lookup table,space spectrum module and the sound source localization module are designed.Finally,each module is simulated by ModelSim simulation and FPGA experiment.According to the simulation results,the FPGA design of each module is correct,and the effectiveness and real-time performance of the sound source positioning system based on MUSIC algorithm are verified by experiments.The innovation point of this thesis is to use FPGA to implement the sound source localization system,which has strong changeability and portability,and strong flexibility.In conclusion,by researching on each part of the sound source localization system in this thesis,the algorithm is realized on FPGA and other parts of the sound source localization are also realized,the simulation of the design results verifies the correctness of the sound source localization system,which provides a technical basis for other extensions in the future.