Abstract
To address the limitations of single-path planning algorithms in dynamic and complex environments, this paper proposes a hybrid planning framework that integrates global and local planning using a combination of Rapidly-exploring Random Tree (RRT) and Dynamic Window Approach (DWA) algorithms. The global planning employs the RRT algorithm to generate a global path, ensuring goal-directed and near-optimal paths through a cost-grid-based search mechanism. The local planning utilizes an improved DWA algorithm, which, under the guidance of the global path, dynamically adjusts the direction and speed of the UAV in real time to avoid unknown obstacles and adapt to dynamic environmental disturbances. The proposed method addresses key limitations of conventional DWA, including local optima and limited adaptability in dynamic environments. The two planning methods are integrated through a strategy of selecting key sub-goal points: when the UAV follows the global path and encounters a local minimum, the sub-goal points guide the UAV to escape from the local minimum and optimize the path quality. Extensive simulation experiments validate that the hybrid framework significantly enhances the UAV’s obstacle avoidance capabilities and task execution efficiency.The address of the ROS-Gazebo experimental project - https://app.theconstruct.ai/rosjects/980801/.The address of MATLAB source code - https://github.com/moxing15/rrt-dwa.