This is Intelligent Robotics and Interactive Systems (IRIS) Lab! Our research focuses include

                   


Recent Updates

Oct 15, 2024
๐•-Twitter

๐Ÿ”ฅ๐Ÿ”ฅ โ€œSkills from YouTube, No Prep!โ€ ๐Ÿ”ฅ๐Ÿ”ฅ Can robots learn skills from YouTube without complex video processing? Our "Language-Model-Driven Bi-level Methodโ€ makes it possible! By chaining VLM & LLM in a bi-level framework, we use the โ€œchain ruleโ€ to guide reward learning directly from video demos. ๐Ÿš€Check out our RL agents mastering skills from their biological counterparts!๐Ÿš€


Check out the preprint. Here is a long demo:




Aug 24, 2024
๐•-Twitter

๐Ÿš€ Can a robotic hand master dexterous manipulation in just 2 minutes? YES! ๐ŸŽ‰ Excited to share our recent work โ€œContactSDFโ€, a physics-inspired representation using signed distance functions (SDFs) for contact-rich manipulation, from geometry to MPC. ๐Ÿ”ฅ Watch a full, uncut video of Allegro hand learning from scratch below! We are pushing the boundaries of โ€œfastโ€ learning and planning in dexterous manipulation.


Check out the webpage, preprint, and code. Here is a long demo:




Aug 19, 2024
๐•-Twitter

Can model-based planning and control rival or even surpass reinforcement learning in challenging dexterous manipulation tasks? Our answer is a resounding YES! PROUD to share: ๐Ÿ”ฅ๐Ÿ”ฅ"Complementarity-Free Multi-Contact Modeling and Optimization,", our latest method that sets shattering benchmarks in various challenging dexterous manipulation tasks.

"Complementarity-Free Multi-Contact Modeling and Optimization," consistently achieves state-of-the-art results across different challenging dexterous manipulation tasks, including fingertip 3D in-air manipulation, TriFinger in-hand manipulation, and Allegro hand on-palm manipulation, all with different objects. Check out the demo below!

Our method sets a new benchmark in dexterous manipulation:
  • ๐ŸŽฏ A 96.5% success rate across all tasks
  • โš™๏ธ High manipulation accuracy: 11ยฐ reorientation error & 7.8 mm position error
  • ๐Ÿš€ Model predictive control running at 50-100 Hz for all tasks

Check out our preprint, and try out our code (fun guaranteed). Here is a long demo:




July 9 2024
๐•-Twitter

๐Ÿค– Robots may be good at inferring a task reward from human feedback, but how about inferring safety boundaries from human feedback? In many cases such as robot feeding and liquid pouring, specifying user-comfortable safety constraints is more challenging than rewards. Our recent work, led by my PhD student Zhixian Xie, shows that this is possible, and can actually be very human-effort efficient! Our method is called Safe MPC Alignment (submitted to T-RO), enabling a robot to learn its control safety constraints with only a small handful of human online corrections!

Importantly, the Safe MPC Alignment is certifiable: providing an upper bound on the total number of human feedback in the case of successful learning of safety constraints, or declaring the misspecification of the hypothesis space, i.e., the true implicit safety constraint cannot be found within the specified hypothesis space.

Check out the project website, preprint, and a breaf introduction vide below.