Home
Center for Autonomy
Enabling high-impact research in autonomous system design
About
The Center for Autonomy (CfA) is Illinois's interdisciplinary hub for autonomy and robotics research and education in the Grainger College of Engineering at University of Illinois Urbana-Champaign. The CfA houses the Master of Engineering in Autonomy and Robotics, manages a number of robotics labs, fosters community, and helps incubate high-impact research endeavors.
From self-driving cars to intelligent robotic assistants to remote surgical systems, autonomous technology will revolutionize the way we live, work, and play. In order to enable this revolution, however, advancements in foundational research and workforce development must first take place to provide assured and certified-safe performance.
Application areas include:
UPCOMING & RECENT EVENTS
What Makes a Robot Safe? The Capability-Confidence-Consequence Principle
Friday, November 6, 2026
2:00pm Central Time
Zoom Event
Speaker: Yiwei Lyu (Texas A&M)
Robots are expected to accomplish tasks safely alongside people and other robots. However, they may compromise safety by demanding an impossible maneuver from a teammate, misinterpreting human intent, or triggering harm beyond their immediate objectives. In this talk, Yiwei will introduce the Capability–Confidence–Consequence (3C) Robot Safety Principle: safe decisions must be grounded in robots’ physical capabilities, calibrated to confidence in their understanding, and informed by downstream consequences. She will begin with enforceable safety in heterogeneous multi-robot systems, where differences in motion capabilities and competing interaction demands can make collective safety requirements physically infeasible. She will show how capability-aware responsibility allocation and recovery-preserving control make safety requirements realizable, and how capability representations support control-informed learning that generalizes to unseen robot models. Next, Yiwei will investigate interactive safety in human–robot settings, where ambiguous human intent can leave robots without sufficient understanding to act safely. She will discuss how active inference and confidence-aware risk adaptation enable robots to clarify intent through interaction and adjust their actions as their understanding of safety-relevant behavior evolves. Finally, Yiwei will explore anticipatory safety in robotic manipulation, asking when a seemingly successful action sets harm in motion. She will present consequence-aware planning methods that use structured imagination of how action-induced disturbances propagate through interactions, enabling robots to anticipate cascading effects and plan beyond immediate task success. Together, these directions put the 3C principle into practice, advancing responsible robot autonomy that respects physical limits, acts judiciously under uncertainty, and anticipates the broader consequences of its actions.