Multi-Robot Motion Planning
Overview
Multi-robot motion planning studies how teams can reach goals, cover areas, or manipulate objects without colliding or blocking one another. Approaches include grid-based multi-agent path finding, roadmaps and sampling-based search, conflict-based coordination, and trajectory optimization with model predictive control. Learning methods such as reinforcement learning, graph neural networks, and diffusion models increasingly help generate or refine plans, often alongside explicit safety checks. A central challenge is scaling to larger, more dynamic teams while keeping plans safe, executable under real robot dynamics, and robust to uncertainty or limited communication.
Latest papers
150 papers27 papers in the last four weeks, up 42% on the four weeks before. 0.2% of all new papers.