基于模糊集合改进的TDOA声源定位方法

A Specific Sound Source Localization Method Based on Improved TDOA

摘要:

面向在复杂场景下对特定的声源进行筛选并定位的情况,设计了一套声源定位系统该系统以STM32为主控芯片,以基于时间差定位算法(the time differences of arrival,TDOA)作为基本原理。对于可能出现的低信噪比、线性度不贴合实际、以及多径效应等情况,首先利用模糊集合对采集到音频数据进行筛选处理处理,增强信号的抗干扰性,从而得到理想的音频数据。之后利用网格迭代法估算声源坐标。在相同条件下与传统TDOA算法进行实验对比,结果表明在对特定频率的音频进行定位时,整个系统相比于传统算法更具有抗干扰性和定位准确度,达成了预期目标。

Abstract:

A sound source localization system was designed for the screening and localization of specific sound sources in some complex scenarios. The system uses STM32 as the main control chip and the time differences of arrival (TDOA) algorithm as the basic principle. For possible situations such as low signal-to-noise ratio, impractical linearity, and multipath effects, fuzzy sets are first used to filter and process the collected audio data to enhance signal anti-interference and obtain ideal audio data. Afterwards, the grid iteration method is used to estimate the sound source coordinates. Under the same conditions, experimental comparisons were made with traditional TDOA algorithms, and the results showed that the entire system had better anti-interference and higher positioning accuracy compared to traditional algorithms when locating specific frequency audio, achieving the expected goals.