| Summary: | In the traditional rapidly exploring random tree (RRT) algorithm, the planned path is not smooth, the distance is long, and the fault tolerance rate of the planned path is low. Disturbances in an environment can cause unmanned surface vessels (USVs) to deviate from their planned path during navigation. Therefore, this paper proposed a path-planning method considering environmental disturbance based on virtual potential field RRT* (VPF-RRT*). First, on the basis of the RRT* algorithm, a VPF-RRT* algorithm is proposed for planning the planning path. Second, an anti-environmental disturbance method based on a deep recurrent neural networks PI (DRNN-PI) controller is proposed to allow the USV to eliminate environmental disturbance and maintain its track along the planning path. Comparative simulation experiments between the proposed algorithm and the other algorithms were conducted within two different experimental scenes. In the path-planning simulation experiment, the VPF-RRT* algorithm had a shorter planning path and a smaller total turning angle when compared to the RRT* algorithm. In the path-tracking simulation experiment, when using the proposed algorithm, the USV could effectively compensate for the impact of environmental disturbance and maintain its navigation along the planning path. In order to avoid the contingency of the experiment and verify the effectiveness and generality of the proposed algorithm, three experiments were conducted. The simulation results verify the effectiveness of the proposed algorithm.
|
|---|