Getting robots to move swiftly and effortlessly through the unstructured world is more challenging than it seems. The RAI Institute is building robots that think, plan, and move like athletes - robots with mobility and intelligence of a professional bike trial rider that can

Learn more about how we’re building and training UMV:

Using reinforcement learning we have expanded the range of techniques the Ultra Mobile Vehicle (UMV) uses to handle terrain and obstacles, including hops, out-of-plane balance, and level-ground flips. Millions of physics-based simulations provide training data to support

Today we’re celebrating three years of the Institute! Huge thanks to our team for a year filled with impactful research. Looking forward to what’s ahead.

Researchers from RAI Institute present Diffuse-CLoC, a new control policy that fuses kinematic motion diffusion models with physics-based control to produce motions that are both physically realistic and precisely controllable. This breakthrough moves us closer to developing

We’ve just opened a free, hands-on robot lab! This summer, the RAI Institute Robot Lab is popping up at CambridgeSide mall. Come see working robots, drive one yourself, and have the opportunity to contribute to robotics research. See our hours and plan your visit —

Our team is at #ICRA2025 this week presenting our latest robotics research! Visit our booth to learn about job opportunities and attend our poster sessions to learn more about our accepted research papers, which are also available on our website —

New tricks loading ...

The control policy tracks and controls retargeted human motion data.  Each maneuver is created with data from about 150 million runs of the simulator and transferred zero-shot to the hardware. This work is part of a collaboration between the RAI Institute and @BostonDynamics .

Reinforcement learning is used to speed the production of behavior for the @BostonDynamics Atlas humanoid robot.  At the heart of the learning process is a physics-based simulator that generates training data for a variety of maneuvers.

Using reinforcement learning, we trained policies for @BostonDynamics Spot that allow the robot to achieve record running speeds of 11.5 mph (5.2 m/s) — over three times faster than Spot's default max speed.

In this demo, Ultra Mobile Vehicle (UMV) drives, turns, jumps, tricks, and comes to a sudden stop called a track-stand. All of the driving, landings, balance, and track-stands are done using reinforcement learning.

“The key of RL is to discover new behavior and make this robust and reliable under conditions that are very hard to model. That’s where RL really, really shines.” Marco Hutter and Farbod Farshidian discuss the latest progress from the Institute. https://t.co/Xc0UKqObK2

Announcing our new partnership with @BostonDynamics! We're teaming up to advance humanoid robotics through reinforcement learning. Learn more —

The AI Institute is now the Robotics and AI Institute, or RAI Institute! Our name and branding have evolved to better reflect who we are and what we're building. Stay tuned for more exciting things ahead.

Introducing Theia, a vision foundation model for robotics developed by our team at the Institute. By using off-the-shelf vision foundation models as a basis, Theia generates rich visual representations for robot policy learning at a lower computation cost. https://t.co/aTy77T5xfT

Our team is presenting work at the Conference on Robot Learning, @corl_conf, in Munich, Germany this week! Learn more about our accepted research —

We’ll be at the Grace Hopper Celebration in Philadelphia next week. If you're attending, stop by kiosk 1212-10 to meet some of our amazing women engineers and learn about opportunities to join our team. We’re hiring across many roles, and we’d love to chat with you!

The next generation of robots is right around the corner. Help us shape the future. Join our team — https://t.co/r4d6chmXLW