Design Of MEMS Acoustic Sensor Array Target Location System

The development of military confrontation exercises places increasing requirement on the intelligence of the target system for simulating combat objects,which requires the target system to be able to simulate all aspects of the intelligence of the training combat object.The "intelligent moving target" needs to be able to perceive the surrounding situation and react appropriately.The ability to simulate human hearing and to locate and judge the "heard" sounds is one of the functions of an "intelligent moving target".In this paper,the design of a MEMS acoustic sensor array target positioning system is completed in accordance with the requirements of the "Intelligent Mobile Targeting System" open cooperation project of the China Research Institute of Weapons Industry Experimentation and Testing.The main work is as follows:（1）The superiority of MEMS acoustic sensor array localisation in intelligent mobile target applications was analysed.Furthermore,the principle analysis and characteristics comparison of beamforming-based methods,High-resolution spectral estimation sound source localization algorithm and Time-delay estimation sound source localization algorithm（TDOA）were conducted,and the TDOA with high real-time performance,small computation was selected as the sound source localization algorithm.Combined with the actual requirements of the project background,the key technical indicators of the MEMS acoustic sensor array target localisation system were determined.（2）The performance of the algorithm for short baseline arrays is simulated.The differences in arrival time difference estimation between standard and generalised mutual correlation calculations,which are easy to implement on embedded systems,are emphatically analysed.The simulation implements BLUE-TDOA-based source localisation and source orientation performance simulations.The results of the simulations show that the arrival time difference calculation based on the standard mutual correlation algorithm can meet the requirements of source orientation and the accuracy of source orientation is achieved in the simulation environment.（3）Spectrum analysis of common sound sources in confrontation drills.The results show that similar spectral features can be obtained using short sound sample data,thus laying the foundation for the next step in the identification of sound source types based on spectral features.（4）The characteristics of three different microphone array configurations were compared.A seven-array module with a hexagonal bottom surface consisting of eight MEMS acoustic sensors MSM261S4030H0 R was designed as a base and extended with one sensor at the top,thus forming a six-sided conical three-dimensional structure of the MEMS acoustic sensor array.The design of the seven-array module on the underside is based on the Sipeed R6+1 MEMS microphone array module from Shenzhen Sipeed Inc;and the 12 SK9822-LED beads on the module are used to indicate the direction of the sound source.（5）The design implemented the hardware and software for the sound source localisation system with STM32 as the core and Nebula＿Io T software resources as the basis.The C language was used to compile the sound acquisition,standard correlation calculation,BLUE-TDOA sound source localisation and directional solving.Verification shows that although the acquisition rate of MEMS acoustic sensor through GPIO port is not enough to achieve synchronous multi-channel data acquisition,but the efficiency is relatively high in correlation calculation and location solution.The solution designed in this paper for TDOA arrival time difference estimation and BLUE source direction estimation can meet the orientation requirements.