A multi-objective bi-level task planning strategy for UUV target visitation in ocean environment

Unmanned underwater vehicle (UUV) is commonly utilized for ocean resource exploration. To effectively plan long-term tasks, it is crucial to consider energy usage and task quality. This paper proposes a multi-objective bilevel task planning strategy (MOBTPS) for solving an UUV dispatched to visit a set of targets. On the one hand, rapid initialization screening method based on task quality is adopted. On the other hand, to address the challenge of black-box optimization for UUV task time, a nested optimization approach is employed. The upper level of optimization focuses on determining the shortest access order for the tasks, while the lower level optimizes the route between the task points. The Simulated Annealing (SA) and Genetic Algorithm (GA) are selected for simultaneous optimization of task assignment and path planning. In order to adapt to varying ocean currents, an UUV control strategy is incorporated into the path planning process. The optimal solution is obtained by using the criteria importance through intercriteria correlation (CRITIC) method. The effectiveness of MOBTPS is demonstrated through extensive numerical simulations.