Intelligent Navigation System for Unmanned Surface Vessel Based on RRT* and SAC

In recent years, researches on unmanned surface vessel have received increasing attention. In order to achieve the secure navigation of intelligent unmanned surface vessels, this paper designs an Intelligent Navigation System (INS) based on RRT* and SAC algorithm. The system can improve the navigation efficiency, and take intelligent decision-making in static and dynamic environment with unknown obstacles. Firstly, based on INS, in the planning module, the rasterized and grayscale processed electronic chart is input to RRT* algorithm for pre global path planning. Secondly, in the decision-making module, the states, actions, and reward functions are set, and the model is trained based on the soft actor critic (SAC) algorithm. Finally, four simulation scenarios are set up to verify the generalization of the model. Experiments have shown that the model based on INS can take effective collision prevention measures in various unknown environments, ensuring the safety of navigation.