An improved A*-DWA mobile robot path planning algorithm considering path smoothing and dynamic obstacle avoidance

Abstract

To reduce redundant search, frequent path turning, and trajectory oscillation in mobile robot navigation, this paper proposes an improved A*-DWA fusion path planning algorithm considering path smoothing and dynamic obstacle avoidance. In the global planning stage, a joint cost function integrating a dynamic heuristic weight, a turning penalty, and an obstacle-distance penalty is introduced into A* to improve search efficiency and path quality. Then, key-point extraction and conservative corner smoothing are adopted to reduce abrupt path turning while maintaining obstacle safety. In the local planning stage, a global path guidance term and an angular velocity smoothing term are added to the DWA evaluation function to improve trajectory tracking stability under dynamic obstacle interference. MATLAB simulations are conducted in simple static, complex static, and dynamic obstacle scenarios. The results show that improved A* reduces the number of search nodes from 238 to 56 in the simple map. Conservative smoothing reduces the cumulative turning angle from 360.00° to 214.21° in the simple map and from 900.00° to 116.21° in the complex map while maintaining collision-free safety. In dynamic obstacle scenarios, the proposed method achieves a 100% arrival success rate and reduces angular velocity variation from 32.2 to 3.9 compared with A*+DWA. These results demonstrate improved global search efficiency, path trackability, and dynamic obstacle avoidance stability.