Detection in present of reverberation combined with blind source separation and beamforming

Reverberation is a main problem to sonar signal processing. Based on the proposed data model of the active sonar, we present a new algorithm for detection in present of reverberation. The proposed method uses the sliding window divide the data into several blocks first, blind source separation (BSS) is used to separates the sonar data into reverberation component and target component, and then the two components is reconstructed to array data, beamforming is used to estimate the direction and enhance the target echo at last. It also used the simulation data and sea trial data to verify the proposed method. Compared with the classical algorithm, the method can get the target location exactly and advance the output signal and reverberation ratio (OSRR) to 4dB at least. The result shows that, the proposed data model is accord to the characteristic of reverberation and the method has better performance for reverberation canceling than the classical method.