Switching target in position-based visual servoing for collision-free underwater docking of an autonomous underwater vehicle

An autonomous underwater vehicle (AUV) can only operate for a limited time due to limited energy supplied by embedded battery. Its operation can be prolonged when the battery is recharged at an underwater docking station. Several researches had been conducted for the docking of AUV using vision. Most of them use image-based visual servoing which focuses solely on features of the docking station to guide the AUV. However, image features are highly non-linear function of camera pose. Furthermore, many collisions occurred during docking operation which may damage the platforms if not docked carefully. In this paper, a switching target in position-based visual servoing is proposed for the AUV to securely dock inside the station. The features of the docking station are extracted using image processing. Then, simulation is conducted using the proposed method showed that the AUV is able to enter the docking station without any collision.