Seamless Underwater Navigation with Limited Doppler Velocity Log Measurements

Abstract

Autonomous Underwater Vehicles (AUVs) commonly utilize an inertial navigation system (INS) and a Doppler velocity log (DVL) for underwater navigation. To that end, their measurements are integrated through a nonlinear filter such as the extended Kalman filter (EKF). The DVL velocity vector estimate depends on retrieving reflections from the seabed, ensuring that at least three out of its four transmitted acoustic beams return successfully. When fewer than three beams are obtained, the DVL cannot provide a velocity update to bind the navigation solution drift. To cope with this challenge, in this paper, we propose a hybrid neural coupled (HNC) approach for seamless AUV navigation in situations of limited DVL measurements. First, we drive an approach to regress two or three missing DVL beams. Then, those beams, together with the measured beams, are incorporated into the EKF. We examined INS/DVL fusion both in loosely and tightly coupled approaches. Our method was trained and evaluated on recorded data from AUV experiments conducted in the Mediterranean Sea on two different occasions. The results illustrate that our proposed method outperforms the baseline loosely and tightly coupled model-based approaches by an average of 96.15%. It also demonstrates superior performance compared to a model-based beam estimator by an average of 12.41% in terms of velocity accuracy for scenarios involving two or three missing beams. Therefore, we demonstrate that our approach offers seamless AUV navigation in situations of limited beam measurements.

Figures (7)

• Figure 1: A visualization of various scenarios where an AUV may encounter conditions limiting DVL beam measurements. On the left-hand side, the initial operation of the AUV, diving, is illustrated, characterized by a significant pitch angle. In the middle, the acoustic beams encounter uneven terrain, while on the right-hand side, the DVL view is obstructed by sea animals.
• Figure 2: A schematic representation of the revised BeamsNet architecture, comprising of one-dimensional convolutional layers and fully connected layers.
• Figure 3: Hybrid-Neural EKF information flow. The EKF is initialized and constantly propagates the INS model while awaiting the DVL velocity update. Upon receiving the DVL velocity update, if the bottom lock condition is met, the update procedure is carried out regularly, either in an LC or TC approach. However, if there are fewer than three beams, the data is directed through the appropriate path and the missing beams are forecasted using the proposed approach and subsequently utilized within the filter.
• Figure 4: Example of two out of the eight trajectories used to train our proposed network.
• Figure 5: Velocity error states. The blue line represents the velocity error state vector bounded within the estimated covariance sleeve (shown in red) of the EKF for trajectory 1 in (a) and trajectory 2 in (b).