Robotics & Physical AI
A robot that can think but not act is a simulation. A robot that can act but not reason is a tool. The work at Rice Computer Science is about building systems that do both in environments that do not cooperate.
Our research focuses on robotic agents that can reason, adapt and act in the physical world.
- The Kavraki Lab works at the intersection of robotics, computational biomedicine and physical AI, producing the Open Motion Planning Library (OMPL), an open-source toolkit used in research labs worldwide.
- The Human-Centered AI and Robotics (HCAIR) Group, led by Vaibhav Unhelkar, focuses on building AI systems that work alongside people: human-AI teaming, decision support and collaborative robotics with applications in healthcare and industry.
- We also investigate algorithm-embodiment co-design: how a robot's physical structure and its algorithms should inform each other.
A system built for a clinical environment needs different reasoning than one built for a factory floor, and both differ from a search-and-rescue system operating in conditions it was never designed for.
Target application areas include manufacturing, clinical healthcare, transportation, search and rescue and human-centered robotic agents that support people in education, independent living and daily tasks.
“Physical AI is not about making robots smarter in simulation. It is about making them reliable in the real world.”
What We Explore
- How does a robot plan a path through an environment it has never encountered before?
- What does it take for a robotic system to interact safely with a person, not just avoid them?
- How do you design a robot arm for a clinical setting where the consequences of failure are serious?
- What is the right balance between autonomy and human oversight in high-stakes systems?
- How does a robot generalize from a few demonstrations to novel tasks it was not trained on?
Faculty
Faculty members leading research in Robotics & Physical AI are as follows:
- Kaiyu Hang NSF CAREER Award (2023) - robot manipulation and dexterous grasping
- Lydia E. Kavraki Kenneth and Audrey Kennedy Professor of Computing; NAE; NAM; NAS; AAAS; ACM Fellow; IEEE Fellow; IEEE Frances E. Allen Medal 2023; Allen Newell Award 2019; Director, Ken Kennedy Institute; Kavraki Lab - Physical AI, Robotics & Biomedicine
- Vaibhav Unhelkar Human-Centered AI and Robotics (HCAIR) Group