Dynamic path planning for multiple unmanned ground vehicles based on improved dynamic window approach

This study proposes an improved dynamic window approach for path planning and collision avoidance of multi-Unmanned Ground Vehicle (UGV) in dynamic environments. The proposed method enhances the Predictive Collision Avoidance Dynamic Window Approach (PCA-DWA) by incorporating the kinematic and geometric models of UGVs. Two nonlinear sampling methods, power function sampling and exponential function sampling, are introduced to replace the traditional linear sampling. Additionally, a priority strategy is implemented to further improve the algorithm’s performance. Simulation results and comparative analysis show that the proposed dynamic window approach successfully achieves path planning and collision avoidance for multiple UGVs. The nonlinear sampling methods offer higher control accuracy compared to traditional linear sampling, and the dynamic window approach with the priority strategy significantly enhances the overall efficiency of path planning. These conclusions validate the effectiveness and superiority of the proposed approach.