At the recently concluded IEEE International Conference on Robotics and Automation (ICRA 2025), the universal dexterous grasping framework D(R,O) Grasp: A Unified Representation of Robot and Object Interaction for Cross-Embodiment Dexterous Grasping, presented by Dr. Lin Shao, Chief Scientist at RoboScience and Assistant Professor at the National University of Singapore, stood out among more than 4,000 submissions and was awarded the “Best Paper Award on Robot Manipulation and Locomotion.”

Manipulation and locomotion are the two principal pillars of the field of embodied intelligence. This award is one of the most prestigious academic honors in robotics, recognizing the year’s most cutting-edge innovations in embodied intelligence. The Shao Lin team’s receipt of this award marks the first time in nearly five years that an Asian institution has won the prize as the primary affiliated institution.

D(R,O) Grasp addresses the challenge of generalizing grasping strategies across different types of dexterous hands by innovatively developing a unified representation of robot–object interactions. This approach overcomes the traditional limitation of “one robot, one strategy” and achieves cross‑ontology generalization—“one brain, multiple forms”—enabling a single AI model to support diverse grasping tasks with various dexterous hands. The video demonstrates grasping operations performed by four distinct dexterous hands: LeapHand, Shadow Lite, XHand, and SoftHand, encompassing 3‑, 4‑, and 5‑finger configurations as well as both soft and rigid designs.

                                                                                                                                                                                             
Project homepage:
https://nus-lins-lab.github.io/drograspweb/

Paper:
https://nus-lins-lab.github.io/drograspweb/static/data/D(R,O)Grasp.pdf