基于FPGA的声学相机音频采集系统设计  

Design of Audio Acquisition System for Acoustic Camera Based on FPGA

作　　者：温锦龙 师蔚[1,2] 胡定玉 廖爱华[1,2] WEN Jinlong;SHI Wei;HU Dingyu;LIAO Aihua(School of Urban Railway Transportation,Shanghai University of Engineering Science,Shanghai 201620,China;Shanghai Engineering Research Center of Railway Noise and Vibration Control,Shanghai 201620,China)

机构地区：[1]上海工程技术大学城市轨道交通学院,上海201620 [2]上海市轨道交通振动与噪声控制技术工程研究中心,上海201620

出　　处：《电子科技》2025年第4期66-72,共7页Electronic Science and Technology

基　　金：国家自然科学基金(12274282);上海市地方院校能力建设项目(20030501000)。

摘　　要：针对基于FPGA(Field Programmable Gate Array)的声学相机音频采集方案存在信号采集效率低下、硬件逻辑复杂度高的问题,文中提出一种高效的基于FPGA声学相机音频采集系统设计方案。在FPGA内部使用级联的高效解码滤波器对多通道PDM(Pulse Density Modulation)音频信号流进行解码以减少硬件计算工作量。FPGA采集多通道MEMS(Micro Electro Mechanical Systems)麦克风的PDM数据流进行位拼接并补码为有符号数,各帧数据包通过AXI4-Stream总线协议在各个级联滤波器中进行流水线操作,并行运行整个解码过程来简化硬件逻辑的复杂性。解码后的数据通过UDP(User Datagram Protocol)协议进行千兆以太网传输来保证实时性。仿真和实验结果验证了所提方案的可行性和有效性,数据解码延时小于200μs,硬件逻辑资源占用率小于60%,数据丢包率为0%,表明所提方案实时性高、稳定性强,可为基于FPGA的声学相机设计提供更多参考。In view of the low signal acquisition efficiency and high hardware logic complexity of the acoustic camera audio acquisition scheme based on FPGA(Field Programmable Gate Array),an efficient acoustic camera audio acquisition system design scheme based on FPGA is proposed in this study.Inside the FPGA,a cascaded and efficient decoding filter is utilized to decode the multi-channel PDM(Pulse Density Modulation)audio signal stream to reduce the workload of hardware calculation.FPGA collects the PDM data stream of multi-channel MEMS(Micro Electro Mechanical Systems)microphones for bit splicing and complementing into signed numbers.Each frame data packet is pipelinized in each cascade filter through AXI4-Stream bus protocol.The entire decoding process runs in parallel to simplify hardware logic complexity.The decoded data is transmitted over Gigabit Ethernet using UDP(User Datagram Protocol)to ensure real-time transmission.Simulation and experimental results verify the feasibility and effectiveness of the proposed scheme.The data decoding delay is less than 200μs,the hardware logic resource occupancy is less than 60%,the data packet loss rate is 0%,which demonstrate that the proposed method has high real-time performance and strong stability and can provide more beneficial enlightenment for the design of FPGA-based acoustic camera.

关 键 词：FPGA 声学相机 PDM音频解码 MEMS麦克风 声源定位 AXI4-Stream总线协议 千兆以太网 滤波器设计 

分 类 号：TP23[自动化与计算机技术—检测技术与自动化装置]