The Local Path Planning Algorithm for Amphibious Robots Based on an Improved Dynamic Window Approach

The autonomous navigation capability of amphibious robots in complex water–land environments is a key technology. However, with the existing local path planning methods, it is difficult to meet the autonomous navigation needs of amphibious robots. To address the shortcomings of unreachable targets, poor adaptability, and limited planning range in a water–land environment, this study proposes a local path planning method based on the improved dynamic window approach (IDWA). A water–land hybrid kinematic model and an obstacle expansion method are applied in the new approach. The improved dynamic window approach enhances the automatic adaptability of complex water–land environments. The improved evaluation function and distance cost function avoid overshooting at the target endpoint. The speed resolution adaptive adjustment algorithm improves the ability to pass through a complex multiple-obstacle area, and the dynamic obstacle prediction algorithm optimizes obstacle avoidance paths. The simulation and lake experiments demonstrate that compared to traditional DWA, the IDWA reduces task completion time by 32.99%, algorithm runtime by 35.29%, and path length by 10.78%. The heading angle variations are decreased by 9.92% while maintaining an average speed of 0.70 m/s in complex environments. The experimental results validate that the proposed approach can effectively plan safe and smooth paths in complex water–land environments with multiple moving obstacles.