LOCATION ESTIMATION OF BROADBAND SOURCE WITH ADAPTED MUSIC ALGORITHM

This study deals with location estimation of the broadband acoustic source in the distributed sensor network. The proposed algorithm is based on the MUSIC (Multiple Signal Classification) algorithm, which solves the localization problem in time domain. The broadband source signal arrives at each sensor with different amplitudes and delay times. Therefore, the signals which are reached at the sensors are the noisy versions of the each other. The collected sensor data is called the observation set, and the cross correlation matrix is calculated by using the observation set. The correlation matrix provides a measure of the similarity between each sensor signal and the relative delay time between the received sensor signals which can be determined by using this data. A reference time point is chosen among the received signals and each time delays are eliminated according to this signal after calculating the cross correlation matrix. Consequently, the source location estimation task is performed by MUSIC algorithm in time domain. Moreover, the simulation results which demonstrate the performance of the method are given in the paper.