Hybrid Path Planning for Mobile Robot Based on Improved A* Fusion Dynamic Window Approach

Mobile robot’s navigation has been especially challenging in dynamic environments. The paper proposes a hybrid path planning to avoid obstacles in grid-map representation. First, global path planning (GPP) is developed by a traditional A* algorithm in the known environment. Then, the A* heuristic function adds one more risk coefficient to ensure the ability to safely avoid obstacles. Besides the task of ensuring the shortest path, unnecessary path nodes are removed to optimize the computation process. The GPP is continuously smoothed by the B-spline method. Furthermore, real-time path planning is improved by dynamic window approach (DWA) in continuous local regions following the GPP. The simulations and practical examples validate the feasibility and performance. Finally, the proposed hybrid path planning deals with the challenges of dynamic obstacle avoidance.