Locating the acoustic source in thin glass plate using low sampling rate data.

Acoustic source localization is an important step for structural health monitoring (SHM). There are many research studies dealing with localization based on high sampling rate data. In this paper, for the first time, acoustic source is localized on an isotropic plate using low sampling rate data. Previous studies have mainly used a cluster of specific sensors to easily record high sampling rate signals containing qualitative time domain features. This paper proposes a novel technique to localize the acoustic source on isotropic plates by simply implementing a combination of two simple electret microphones and Loci of k-Tuple Distances (LkTD) from the two sensors with low sampling rate data. In fact the method proposes substitution of previous methods based on solving the system of equations and increasing the number of sensors by implementing the selection of LkTD. Unlike most previous studies, estimation of time difference of arrival (TDOA) is based on the frequency properties of the signal rather than it's time properties. An experimental set-up is prepared and experiments are conducted to validate the proposed technique by prediction of the acoustic source location. The experimental results show that TDOA estimations based on low sampling rate data can produce more accurate predictions in comparison with previous studies. It is also shown that the selection of LkTD on the plate has noticeable effects on the performance of this technique.