基于Raspberry Pi的声源定位跟踪系统

Raspberry Pi Based Sound Source Location Tracking System

研究设计介绍了以Raspberry Pi开发板为核心处理器,配备麦克风阵列,开发版为Arduino板等组成的一个声源定位跟踪系统。系统基本工作流程为:由麦克风阵列拾取语音信号,继而通过Raspberry Pi开发版对信号的处理,其中信号处理方法包括FFT算法滤波降噪以便更精准的接收声源,再通过TDOA算法得到时差数据信息,进一步确定声源位置,获取位置后通过串口传输给二维云台,二维云台承载激光笔转向至声源处并开启激光指向声源,当声源移动时,云台实现动态追踪。其次,基于到达时间延时的定位算法很多,改进了一种的互功率谱相位加权算法。该算法运算量少,经实验证明,在强噪声环境下该算法具有较好的定位性能。

Introduces a sound source location tracking system based on the Raspberry Pi development board as the core processor,equipped with a microphone array and an Arduino board as the development version.The basic workflow of the system is as follows:the microphone array picks up the voice signal,and then the signal is processed by the Raspberry Pi development board,where the signal processing method includes the FFT algorithm filtering and noise reduction for more accurate reception of the sound source,and then the TDOA algorithm gets the time difference data information to further determine the location of the sound source,after obtaining the location,it is transmitted to the 2D head through the serial port,and the 2D head carries The 2D head carries the laser pointer to the sound source and turns on the laser to point at the sound source,when the sound source moves,the head realizes dynamic tracking.Secondly,there are many localization algorithms based on time delay of arrival,and an improved mutual power spectrum phase weighting algorithm.The algorithm is less computationally intensive and has been experimentally demonstrated to have better localization performance in a strong noise environment.