Abstract
A methodical approach for evaluating each sound intensity from the vehicle stream is described. By the Fourier transform theory, each frequency component of sampled sound intensity by moving n point sources can be expressed by a linear equation of n complex exponentials. Applying the Prony's method to above equation, both vehicle velocity and time when every vehicle passes at the front of the observation point can be determined by the roots of polynomial equation. And each sound intensity can be chosen in such a way that the sum of squared norm is minimized. For the actual data, the total intensity reduction from point source is assumed to be expressed by a product of the attenuation by the distance and cosine function with which the excess attenuation is represented simply. An experimental results shows that the recomposed sound intensity by the estimated results gives good approximation to actual noise level tracing. This method can be applicable to identify the fixed noise sources by using traveling microphone, and also has general applicability in the field where inverse-square-law exists.
