Local path planning of under-actuated AUV based on VADWA considering dynamic model.

Dynamic window approach (DWA) considering dynamic performance provides good results in local path planning of autonomous underwater vehicle (AUV). However, the path planned by DWA sometimes detours, and the dynamic performance is limited by constants unreasonably. In this paper, an improved method called vector field adaptive DWA (VADWA) is proposed. Firstly, adaptive parameters and vector field histogram are introduced to solve the problem of routing detouring, so the evaluation parameters and evaluation system of DWA are both optimized, which significantly improves the obstacle avoidance efficiency of AUV. Secondly, different from the practice of simplifying the dynamic model to constants in DWA, VADWA considers the real dynamic model and the input characteristics of actuator to evaluate the dynamic performance of AUV, so as to give more reasonable planning signals. Finally, aiming at the problem of the desired state tracking of AUV in complex environment, a velocity controller is designed to form a complete hierarchical planning-control scheme, which realizes the stable and accurate tracking of desired state. Simulation results show that VADWA can effectively shorten the distance and time for AUV to reach the target point, and generate control instructions that are more consistent with dynamic characteristics of AUV. • An adaptive dynamic window approach is proposed to improve the flexibility of AUV local path planning. • The vector field histogram is introduced into the dynamic window approach to optimize the evaluation system of the algorithm. • Considering dynamic model of AUV and characteristics of actuator, the dynamic performance of AUV is evaluated dynamically. • A complete hierarchical planning-control scheme of AUV is formed, which greatly facilitates AUV velocity tracking. [ABSTRACT FROM AUTHOR]