RAI Institute

Dynamic manipulation steps up to the plate! This is a first look at a low-impedance platform designed to study how robots manipulate objects. In this demo, two robots play catch and practice batting, even teaming up with humans. The robots are capable of throwing 70mph [112 kph] and catching and batting at short distances (23ft [7m]). This requires quick reaction times to catch balls thrown at up to 41mph [66kph] and hit balls pitched at up to 30mph [48kph]: https://lnkd.in/gdUWrPSQ

We’re at the AnitaB.org Grace Hopper Celebration in Chicago this week! If you’re attending, make sure to add our panel discussion on Multiple Career Paths to Robotics to your schedule: https://lnkd.in/eNikejSk And while you’re exploring the expo halls, stop by our booth #314 to learn more about the research in robotics and AI we’re doing at the RAI Institute. #GHC25 #WomenInTech

Wishing everyone a safe and not-too-scar-RAI Halloween!

See Spot perform dynamic whole-body manipulation. Using a combination of reinforcement learning (RL) and sampling-based control, the robot is able to autonomously drag, roll, and stack tires weighing 15 kg (33 lb), well above its maximum arm lift capacity. Learn more about coordinating locomotion and manipulation processes: https://lnkd.in/gzWQAMuk

Current robotic systems face constraints in loco-manipulation due to task-specific designs and fixed joint configurations, limiting adaptability in diverse environments. The ReLIC (Reinforcement Learning for Interlimb Coordination) framework being presented at CoRL 2025 addresses these challenges through reinforcement learning, enabling flexible interlimb coordination. This system removes traditional locomotion/manipulation barriers, allowing dynamic role switching during task execution. Learn more about this research from The RAI Institute: https://lnkd.in/es2KcNht

Our Executive Director Marc Raibert is one of the first guests interviewed by Brian Heater on his newly launched Automated Podcast by A3. Listen now to hear about Marc’s lifelong love of building robots, how the robotics industry has changed over the last 50 years, and how the RAI Institute is focused on technical progress and problem solving for the next generation of robotics: https://lnkd.in/eXj6Kn-q

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 perceive, plan, and navigate any terrain, no matter how complicated. Read more about how the Ultra Mobile Vehicle (UMV) project uses reinforcement learning (RL) to train the robot and unlocks its physical capabilities: https://lnkd.in/drz_EYB6

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 zero-shot transfers. Learn more about how we’re building and training UMV: https://lnkd.in/eA2Jk6dc

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. 

Today at SIGGRAPH, researchers from RAI Institute are presenting 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 generalist policies that enable humanoid robots to perform diverse tasks, including dynamic locomotion and contact-rich manipulation, in a natural-looking and robust way. Learn more at https://lnkd.in/eDEb92az