Safe Motion Planning

Safe motion planning studies how robots and vehicles can reach goals without collisions while respecting limits on their motion and the environment. Common approaches include control barrier functions, reachability analysis, and model predictive control, alongside graph search, sampling-based planning, and optimization. Learning-based planners increasingly use diffusion models, world models, and risk predictions, often paired with safety filters or formal checks; they are tested in settings such as autonomous driving, UAV flight, manipulation, and multi-robot navigation. A central challenge is balancing reliable safety guarantees under uncertain perception and dynamics with plans that are efficient, responsive, and not overly conservative.