AUV navigation with seabed acoustic sensing*

Abstract

Autonomous Underwater Vehicle (AUV) being a powerful tool for exploring and investigating ocean resources can be used in a large variety of oceanographic, industry and defense applications. AUV navigation is still a challenging task and it is one of the fundamental elements in the modern robotics, because the ability of AUV to correctly understand its position and attitude within the underwater environment is determinant for success in different applications. Due to the absence of external reference sources, AUV navigation is usually based only on the information obtained from Doppler Velocity Loggers (DVL), Inertial Navigation Systems (INS), etc. But this type of navigation is subjected to a continuously growing error because of the absence of absolute position measurements (for example, received from the GPS or GLONASS). These measurements might be provided by observation of so-called feature points like in the case of the Unmanned Aerial Vehicles (UAV). But the big difference between acoustical and optical images makes this problem much more difficult in the AUV case, and to solve it one needs the detailed preliminary mapping of the operational seabed area. The modern advances in the acoustic imaging give rise to AUV navigation approaches based on the absolute velocity measurements. The one we propose in the present paper is analogous to the optical flow techniques for UAV navigation. It is based on the extraction of information related to the AUV absolute motion from seabed map evolution measurements. The principal advantage of the proposed method is that the fusion of the acoustic mapping and the INS data makes it possible to estimate the absolute velocity of the vehicle with respect to the seabed. In this sense the suggested method is close to the multi-beam DVL measurement, but it is based on another physical principles and thus operates better in different environment. While DVL by design operates perfectly over the flat surface [1], the appropriate environment for the suggested method implicates the seabed relief, because it extracts the velocity information from the evolution of the measured distance between the sensor and the seabed.