Dynamic Obstacle Avoidance
Overview
Dynamic obstacle avoidance studies how robots and autonomous vehicles choose feasible paths and actions when people, vehicles, or other obstacles move. Methods range from space-time search and model-predictive control to velocity-obstacle methods and control barrier functions that constrain reactive controllers. Learning-based approaches use reinforcement or imitation learning, motion prediction, and learned risk or occupancy representations, often combined with classical planners and safety filters. A central challenge is balancing safety and efficient progress under uncertain, delayed perception and predictions; reliable real-world transfer and evaluation in dense, interactive settings remain open concerns.
Latest papers
119 papers22 papers in the last four weeks, up 47% on the four weeks before. 0.1% of all new papers.