Template Tracking and Visual Servoing for Alignment Tasks with Autonomous Underwater Vehicles

Alignment of underwater vehicles with respect to underwater structures is of utmost importance for many tasks performed in marine environments, specially for those requiring manipulation and, therefore, a specific and stable vehicle pose. These skills are crucial in the case of autonomous vehicles, where there is not a human in the loop in charge of tele-operating the vehicle to a suitable position and attitude with respect to the target. In this paper we apply 2D visual servoing techniques based on template tracking to the problem of autonomously reaching and keeping a vehicle pose relative to a target in a typical underwater intervention scenario. Assuming that the object of interest has been already detected, we develop a method that autonomously aligns the vehicle relative to it, and keeps the pose during the task. The method is validated, first in simulation, and then in a water tank with the Girona 500 vehicle