Learning and Interactive Robots

The Learning and Interactive Robots (LaIR) group develops intelligent robot learning methods that enable robots to operate safely, adaptively, and efficiently in complex real-world environments. We focus on learning for advanced robot-environment interactions, physical human-robot collaboration, and learning through human guidance, addressing fundamental challenges such as uncertainty, force-based manipulation, dynamic interactions, and limited training data. By combining data-driven learning with human expertise, we aim to create robots that can learn from both experience and user feedback, continuously improving their performance while remaining safe and reliable. Our research spans applications in manufacturing, agriculture, healthcare, logistics, autonomous systems, and household assistance, with the long-term vision of making robots capable collaborators that can learn, adapt, and co-evolve with the people they assist.