A Hybrid Path Planning Strategy of Autonomous Underwater Vehicles
Xinyu Jian, Ting Zou, Andrew Vardy, Neil Bose
Abstract
Autonomous Underwater Vehicles (AUVs) play a unique role in many applications, including oceanographic research, country defense, ecosystem monitoring, to name a few. The autonomy of AUVs is utilized in automatically planning a feasible path/trajectory to a goal point. A robust planner of AUVs should be able to search a collision-free path/trajectory not only in a large-scale known static environment, but also in the environment with dynamic obstacles. This paper demonstrates a modified and combined Dynamic Window Approach (DWA) and Rapidly-exploring Random Tree (RRT*) to plan both the local trajectory and the global path for AUVs in environments where dynamic obstacles may appear. In case of dynamic obstacles, the planner automatically judges the risk of collision and switches from RRT* to DWA if necessary. Then the planner switches back after collision risk is dismissed. Hence, by switching between two algorithms, the balance of real-time computation and the globally optimal solution is achieved. The effectiveness of the proposed hybrid planning strategy is verified by simulation.