Dynamic Goal-Based Adaptive Line of Sight: An X300 AUV Case-Study

Path-following control is a well-known problem in autonomous vehicle guidance, especially in aerial and marine contexts. The popular Line-of-Sight guidance algorithm specifies heading changes according to a constant lookahead distance, which efficiently reduces cross-track errors. Yet, its performance deteriorates in dynamic contexts with sharp turns, high vehicle speeds, or external disturbances such as ocean currents. To alleviate these limitations, Adaptive Line-of-Sight (ALOS) methods have been formulated to compensate for unknown disturbances, yet they still rely on a fixed lookahead distance, jeopardizing their effectiveness in complex three-dimensional navigation scenarios. This paper presents a Dynamic Goal-Based ALOS method enhancing the standard solution through a real-time adaptive lookahead distance mechanism. In contrast to current approaches, our method adaptively tunes the lookahead distance according to cross-track and vertical-track errors and the path curvature for better tracking performance in high-curvature 3D settings. Numerical tests were carried out in a ROS2-based simulation with the X300 Autonomous Underwater Vehicle. The results show that the new method minimizes both cross-track and vertical-track errors over fixed lookahead approaches, especially for difficult serpentine trajectories.