Research And Design Of Microphone Array Acquisition System Based On DSP And FPGA

Microphone array acquisition system plays a fundamental role in acoustic imaging and an important application in noise control,fault diagnosis and low noise equipment development.Entrusted by the Institute of Acoustics,Chinese Academy of Sciences,this thesis studies and designs a sound imaging system based on DSP and FPGA.This thesis firstly explains the acquisition technology of microphone array,introduces several important imaging algorithms,and compares their application occasions,advantages,and disadvantages.Finally,the beamforming algorithm is chosen as the core algorithm,and a corresponding simulation analysis is carried out.To satisfy some specific requirements proposed by the client,this thesis design an equipment with the following components:an array of low-noise MEMS microphone sensors is used to collect external sound signals;high precision ADC chip ADI7768 is used to conduct synchronous sampling and conversion of 64 channels of audio signals;low-cost,high-speed FPGA chip EP4CE10F17C7N is used to design the corresponding data interface to receive and cache a large amount of converted data,high performance,low power DSP chip TMS320C6678 is used to read and image the collected data;the fast Ethernet PHY controller 88E1111 is adopted to realize real-time data upload,so as to ensure non-blocking transmission of large traffic data.This thesis analyzes the hardware system according to the design requirements and completes the main device selection.According to the analysis and selection results,the hardware system design,including schematic diagram and PCB drawing.And according to the drawing of the schematic diagram,the program part of the design.This thesis adopts simulation of common beamforming algorithms on MATLAB,analyzes their advantages and disadvantages.Then,the LCMV algorithm is selected for improvement,and the improved LCMV algorithm,FIR digital filter and time selected base 2FFT are simulated by MATLAB to verify its performance.The results show that the performance of digital filter,fast Fourier transform and beamforming algorithm all meet the design requirements.