Hybrid-Driven Motion Planning for a Bionic Underwater Vehicle With Gliding Ability

This study explores an efficient hybrid-driven motion planning method for a bionic underwater vehicle equipped with gliding capabilities, aimed at enhancing the efficiency of autonomous docking while ensuring reliable visual navigation. To achieve this, we first construct and analyze an energy-time consumption model along with a visual navigation stability parameter. These frameworks assess the total energy and time expended, as well as the likelihood of target loss during visual navigation. We then develop a hybrid-driven switching scheme that alternates between two motion modes to approach and dock with a stationary target. A novel motion planning method, based on the deep Q network (DQN) algorithm, is proposed to identify the optimal switching points within this scheme. This method aims to minimize energy and time consumption while maintaining visual navigation stability. The efficacy of the proposed method is validated through both simulation and controlled pool experiments. Results indicate that this hybrid-driven approach not only stabilizes visual navigation but also improves the efficiency of autonomous docking when compared to single-mode strategies. This research contributes to the advancement of hybrid-driven autonomous vehicles and offers new perspectives on the motion planning and control of underwater vehicles with gliding capabilities.