Local Path Planning for USVs Based on An Improved DWA Algorithm

To enhance the path planning performance of an unmanned surface vehicles (USV), a modified Dynamic Window Approach (DWA) is proposed. This modification aims to overcome issues such as local best solution and poor dynamic collision avoidance performance. Firstly, to improve the safety of USV operation in real environments, an obstacle inflation model is introduced. Secondly, to enhance the USV’s ability to navigate around obstacles, a virtual target point evaluation function is added to the traditional DWA. Finally, to improve dynamic collision avoidance capability and shorten collision avoidance time, the speed sampling space is enhanced by eliminating collision trajectories. Simulation experiments demonstrate that the enhanced DWA exhibits satisfactory collision avoidance capabilities compared to the traditional method in both static and dynamic environments. The improved DWA algorithm is more effective in path planning in complex Marine environment.