Toward Safe Control under Uncertainty: Adaptation and Prediction for Control Barrier Functions

As humans, we walk through life making decisions purposefully or otherwise one step at a time. For many of us, these decisions are solutions to unconscious optimization problems that trade off enrichment (i.e. climb that mountain) with safety (don’t fall off the cliff). Even in the face of staggering uncertainty (weather, wildlife, etc.) we often manage to reach the summit and still make it home safely at the end of the day. How is this possible?

It is through my research on control design for autonomous systems that I seek to answer this question. Though primarily focused on applications of control barrier functions in the context of adaptive and predictive control for safety-critical systems, my work is motivated by the problem of safe control synthesis for uncertain systems in pursuit of satisfying some nominal goal specification. And along those lines, as much as there still is to teach the robotic mind, it may be able to teach us even more about ourselves.