Adam Thorpe is a postdoctoral researcher in the ODEN Institute for Computational Engineering & Sciences at the University of Texas at Austin, working in the area of data-driven methods for control. He received his Ph.D. from the University of New Mexico in 2023. He is broadly interested in developing techniques to solve complex problems in modeling, optimization, and control using statistical learning approaches. His research explores the development of responsive and adaptable human-centered autonomy in cyber-physical systems, utilizing data-driven tools and mathematical frameworks to address the dynamic, heterogeneous needs and complex behaviors of users.
In cyber-physical systems (CPS), the complex dynamics of human behaviors and preferences pose challenges to traditional notions of reliability and safety. This complexity raises questions about how autonomy should operate around humans. My research seeks to answer: How can we design human-centered autonomy for CPS that is responsive, robust, and capable of rapid adaptation to new, unseen contexts? I develop a framework for human-in-the-loop control using tools from functional analysis, designed to be responsive to the heterogeneous, dynamic, and time-varying needs and preferences of individual users. These challenges include the inherently limited nature of human data, necessitating innovative approaches and methods grounded in a principled, mathematical framework crafted to adapt quickly and operate in the low-data regime. I aim to develop naturalistic, context-aware solutions that operate with minimal data and adapt swiftly to new information.