Identification of moving acoustic sources from the pass-by noise

Abstract

The identification of a moving noise source such as a transportation system is usually difficult at high speed since microphones attached to the moving noise source may induce a high level of aerodynamic noise, while the measurement using a fixed microphone suffers from the distortion of the source characteristics due to its motion. In this study a time domain technique based on the finite difference method and regularization is developed for the recovery of the stationary source characteristics from the distorted signal measured at a fixed point. The sources are assumed to be point sources with stationary frequency characteristics, and move along a line at a constant subsonic speed. In the previous work of the present authors recovery of the stationary source signal from a point source requires the information about time origin. For multiple sources it is impossible to set a single time origin that a source signal recovery method using multiple time origins are developed. In the numerical simulations the effects of estimation errors in the parameters, such as the speed of source, time origin, and background noise level, are investigated and the results show the usefulness of this method.