An effective framework for underwater acoustic data acquisition

Undersea sensor networks provide a continuous, remote recording of the underwater acoustic signal, and environmental monitoring with minimal human intervention. Thus, large volumes of acoustic data result to the heavy burden of the communications among the undersea sensor networks, and consume a large of energy. To ameliorate the above problems, in this study, we aim to compress the recordings of the underwater acoustic signal and propose an effective framework for underwater acoustic data acquisition from two aspects: First, an optimal measurement is designed to obtain the adequate compression matrix via shrinking singular value decomposition (SVD) method. Second, an accelerated approximation ℓ₀ norm constraint method is derived to exploit the sparse structure of collected underwater acoustic signal in discrete cosine transform (DCT) domain. By compressing the collected underwater acoustic signal, the proposed methods would reduce both energy consumption in mobile assets and collisions between transmissions made by different sensors. Simulation and at-sea experiment results confirm the effectiveness of the proposed methods in terms of normalized mean square error, structural similarity index and running time.