Underwater object localization using compressive sensing

In this paper, we propose a technique to localize a moving submerged object with a sparse measurement using a hydrophone array. We used a natural muddy pond bottom with 9 m length, 5.5 m width, and 2 m depth. First, we measured the pond impulse response to identify the sound propagation characteristics. In the second phase namely static direction of arrival (DOA), the underwater speaker was placed at half-circle positions with respect to the array to simulate the underwater object trajectory, 0 deg, 30 deg, 60 deg, 90 deg, 120 deg, 150 deg, and 180 deg, respectively. The sound generated by a four-blade propeller submarine toy was measured to track its trajectory as a dynamic DOA measurement. The measured data were transformed into discrete cosine transform coefficients and reconstructed sparsely by using the basis pursuit algorithm. Based on the estimated incident angle and time delay, the reconstructed measurements were then compared to full dictionary measurement. The results showed that the error angle of the origin deviated about 2.5 deg with half of the measured data were lost. This may suggest that the sound propagation was degraded due to bottom reverberation and scatterers. Currently, we do a similar measurement in an open shallow water environment to eliminate the reverberation.In this paper, we propose a technique to localize a moving submerged object with a sparse measurement using a hydrophone array. We used a natural muddy pond bottom with 9 m length, 5.5 m width, and 2 m depth. First, we measured the pond impulse response to identify the sound propagation characteristics. In the second phase namely static direction of arrival (DOA), the underwater speaker was placed at half-circle positions with respect to the array to simulate the underwater object trajectory, 0 deg, 30 deg, 60 deg, 90 deg, 120 deg, 150 deg, and 180 deg, respectively. The sound generated by a four-blade propeller submarine toy was measured to track its trajectory as a dynamic DOA measurement. The measured data were transformed into discrete cosine transform coefficients and reconstructed sparsely by using the basis pursuit algorithm. Based on the estimated incident angle and time delay, the reconstructed measurements were then compared to full dictionary measurement. The results showed that the error angle...