An Improved Dynamic Window Approach with Environment Awareness for Local Obstacle Avoidance of Mobile Robots

Abstract

Local obstacle avoidance is critical for mobile robot<br> navigation. It is a challenging task to ensure path optimality and<br> safety in cluttered environments. We proposed an Environment<br> Aware Dynamic Window Approach in this paper to cope with<br> the issue. The method integrates environment characterization into<br> Dynamic Window Approach (DWA). Two strategies are proposed<br> in order to achieve the integration. The local goal strategy guides<br> the robot to move through openings before approaching the final<br> goal, which solves the local minima problem in DWA. The adaptive<br> control strategy endows the robot to adjust its state according<br> to the environment, which addresses path safety compared with<br> DWA. Besides, the evaluation shows that the path generated from<br> the proposed algorithm is safer and smoother compared with<br> state-of-the-art algorithms.