Moving Point Sound Source Localization Method Based on Time Domain Characteristics of Trackside Acoustic Signals

Abstract:

To investigate the relationship between the characteristics of trackside noise and the geometric defects of the wheel tread in subways， it is necessary to accurately locate the spatial position of the wheel-rail noise source of the moving train. A noise source localization method based on the coherence of trackside signal environment is proposed. First， acoustic signals are collected by deploying sensor arrays. Then， the time difference of arrival （TDOA） calculation method for trackside acoustic time-domain signals was optimized using the Euclidean norm. Finally， the feasibility of this improved moving sound source localization method was verified by comparing it with traditional methods. The results demonstrate that the proposed method achieves better error stability at speeds below 60 km·h^－1. Regarding the localization range， the method performs well within a predicted angle range of θ < 59.5°. The proposed method achieves high-precision localization of moving point sound sources for signals with a signal-to-noise ratio （SNR） above 40 dB.