Robotic Tool Use

Robotic tool use studies how machines choose and control tools to extend what they can do, from cutting and grinding to drilling and household tasks. Methods combine visual affordance grounding, geometric and physics-based planning, tactile or force feedback, and controllers learned from demonstrations, reinforcement learning, or simulation. Research increasingly targets dexterous and articulated tools, transfer to unseen objects, and long-horizon execution by humanoid and other robots. A central challenge is turning correct tool selection into reliable contact-rich action despite changes in tools and materials, imperfect sensing, and sim-to-real gaps.