Experimental comparison of bearing estimation techniques for short passive towed sonar arrays

Although Low Frequency Active Sonar (LFAS) is considered one of the best means of detection and localization of a submarine in shallow water, the use of an active system is not always possible for tactical reasons. It is therefore interesting to include a passive mode in these systems by using the receiving array as a passive array. However, those arrays are smaller than conventional passive towed arrays and the frequencies used for passive submarine detection are lower than the LFAS frequencies, resulting in reduced bearing resolution and accuracy compared to dedicated passive systems. Several signal processing techniques have been studied to improve the performance of such systems in terms of bearing localization. The Extended Towed Array Measurement (ETAM) algorithm is a Synthetic Aperture technique based on the correlation of data snapshots on overlapping hydrophones. It creates a virtual longer aperture that offers better bearing resolution power than a conventional technique. Another widely considered approach is the Minimum Variance Distortionless Response beamformer (MVDR). We have implemented both algorithms and applied them to simulated as well as experimental data sets. In this paper, several passive experiments that were part of a larger Netherlands trial devoted to multistatic Low Frequency Active Sonar (LFAS) are analyzed. The receiving array aperture was 24 m. A target submarine was fitted with a broadband low frequency projector transmitting five intermittent tonals in the 500 Hz-700 Hz band. The bearing estimation accuracy and beamforming gain of MVDR, ETAM and of the conventional beamformer are measured on this data set and compared.