Skip to content

Robotics & Embodied AI

AI that senses and acts in the physical world through bodies — robots, drones, and vehicles.

Most AI lives inside a screen — it reads text or looks at pictures and answers back. Embodied AI is different: it lives inside a body, like a robot, drone, or self-driving car, so it has to sense the real world and act in it. Think about catching a ball. Your eyes track where it's going, your brain predicts where it will land, and your arm and hand move to meet it — all in a split second, adjusting the whole way. A robot faces the same challenge: it uses cameras and other sensors to see, decides what to do, and sends signals to motors to move. The hard part is that the real world is messy and unforgiving — floors are slippery, objects are heavy, and a wrong move can knock something over. Getting a machine to handle all that gracefully is one of AI's biggest challenges.

The main ideas

  • Perception — Fusing cameras, lidar, and other sensors into a model of the environment.
  • Control — Turning decisions into motion — classical control through learned policies.
  • Manipulation & locomotion — Grasping, moving objects, walking, and dexterous tasks.
  • Sim-to-real — Training in simulation and transferring to real hardware.
  • Autonomous vehicles — Self-driving stacks: perception, prediction, planning, and control.

Computer Vision · Reinforcement Learning

Want to make things?

Head to AI School — AI camps where kids build their own games.