Timezone: »

Certifiably-correct Control Policies for Safe Learning and Adaptation in Assistive Robotics
Keyvan Majd · GEOFFEY CLARK · Tanmay Khandait · Heni Ben Amor

Fri Dec 09 04:15 PM -- 04:30 PM (PST) @

Author Information

Keyvan Majd (Arizona State University)

I am currently a PhD student in the School of Computing and Augmented Intelligence at Arizona State University, co-advised by Dr. Heni Ben Amor and Dr. Georgios Fainekos. I graduated my M.Sc. in Electrical Engineering from North Carolina A&T State University at 2019. Before that, I received my B.Sc. in Electrical Engineering from Ferdowsi University of Mashhad at 2015. The primary focus of my research is to enable robots to safely operate in the real world while interacting with humans, in particular in assistive robotics domain. I approach this by combining the strengths of planning and control with data-driven techniques. Specifically, my research topic includes integrating the global optimization techniques with reinfocement learning and imitation learning to ensure the robot follows a set of safety specifications. To achieve this, I am developing a novel framework that combines imitation learning and verification to learn safe policies for robots. I am also interested in developing computationally efficient optimization-based techniques for learning safe policies for robots.

GEOFFEY CLARK (Arizona State University)
Tanmay Khandait (Arizona State University)
Heni Ben Amor (Arizona State University)

More from the Same Authors