BeamsNet: A data-driven approach enhancing Doppler velocity log measurements for autonomous underwater vehicle navigation

Autonomous underwater vehicles (AUV) perform various applications such as seafloor mapping and structure health monitoring. Commonly, an inertial navigation system aided by a Doppler velocity log (DVL) is used to provide the vehicle’s solution. In fusion, DVL provides vector of AUV, which determines solution’s accuracy helps estimate states. This paper proposes BeamsNet, end-to-end deep learning framework regress estimated that improves estimate, could replace model-based approach. Two versions differing in their input network, are suggested. The first uses current beam measurements sensor data, while other utilizes only taking past for regression process. Both simulation sea experiments were made validate proposed approach relative Sea with Snapir AUV Mediterranean Sea, collecting approximately four half hours data. Our results show achieved improvement more than 60% estimating vector.