Abstract: In order to eliminate the effects of the ocean environmental uncertainty, the accidental errors due to acoustic signal receiving and processing on positioning accuracy, the localization of a moving sound source is realized using EKF method based on the time difference of arrival between the first and second arrival waves. The acoustic waves arriving from multiple paths are identified using Butterworth band pass filtering method and Bellhop ray model, and their relative energies are also compared. And the first and second arrival waves with bigger signal-to-noise ratio (SNR) are studied. The time difference of arrival between the first and second arrival waves is calculated using time domain correlation method by taking single frequency signal as a matched sequence. The time difference of arrival is set to be a measured value, and the acoustic source is localized based on extended Kalman filter (EKF) method, thus obtaining the depth, horizontal range and speed of the moving acoustic source. When the acoustic source and hydrophone are on the surface layer in deep water acoustic channel, the direct and reflected waves overlap each other, and the energies of the first arrival wave and the first wave reflected from the bottom of the sea are higher than the others. The research work shows that the calculated results are in good agreement with the experiment data, which proves that EKF method is well fit for the localization of the underwater moving acoustic source. Key

Key words: localizationofmovingacousticsource, extendedKalmanfilter, timedifferenceofarrival, timedomaincorrelation, multipatharrivalwave