Dr. Nadia Figueroa, a new GRASP faculty member at the University of Pennsylvania, presents her research on collaborative, human-aware robotics and the challenges of making robots learn, adapt, and work safely with people.
Figueroa studies robots that adapt as people move while jointly handling objects. She places this at the most interactive end of a collaboration spectrum, beyond separate workspaces or taking turns. Examples include moving large objects in factories and warehouses, as well as assisting with repetitive household tasks.
Figueroa identifies efficient teaching, adaptation, and safety as core objectives. Robots should learn a task with little human effort, adjust when conditions change, and avoid collisions with people, themselves, and the environment. She argues that human-centered spaces are unpredictable, making these capabilities essential.
Her approach represents robot movement as a dynamical system and learns it from demonstrations, while stability constraints guide the learned behavior. Because the controller responds to the robot’s current state, it can adjust to disturbances, including a person redirecting a robot or two arms tracking a fast-moving object. She describes applications across robot arms, humanoids, wheelchairs, and multi-robot systems.
Figueroa says real deployments must handle unreliable perception, including objects that disappear behind people and later reappear. She also distinguishes motion-level adjustments from task-level understanding: a robot moving beans between bowls needs to detect whether the beans were dropped, not just follow its learned path. In another example, a camera visibility requirement lets a robot return to scanning when a pushed cube leaves the expected route.
Figueroa wants multiple people and robots to coordinate when moving large objects through tight spaces, where understanding the intended action matters as much as changing motion. She summarizes her research priorities as safe interaction, adaptation at motion and task levels, and efficient learning. Potential application areas include industry, homes, and rehabilitation robotics.
Briefs summarise talks and interviews; timestamps open the video at that moment. Claims in a talk are the speaker's own.
A GRASP Lab seminar host and panelists Nadia Figueroa, Dinesh Jayaraman, and Mark Yim discuss robotics research, deployment challenges, and future opportunities at the University of Pennsylvania.
