Infinite Dimensional Optimization for Safety Critical Human-in-the-Loop Systems
A predominant portion of healthcare spending is devoted to the medical care of unintentional injuries, such as those arising from car accidents or falls. By incorporating automation to predict the likelihood of injury and to design and verify personalized treatment, the burden on healthcare professionals, and thus the overall cost of treatment, can be greatly reduced. This talk describes several new techniques each relying upon a theoretical framework grounded in infinite dimensional optimization to improve automation in human-in-the-loop systems.
The first technique is a provably convergent switched system optimal control algorithm to automatically identify an individual-specific dynamic model of locomotion. The second technique is a semi-autonomous architecture that constructs a real-time driver-specific model, which informs a controller that is able to safely correct dangerous driver input to prevent vehicular accidents. The final technique is a scalable convex programming approach for simultaneous reachable set computation and personalized controller synthesis for safety critical applications.
Ram Vasudevan is an assistant professor in Mechanical Engineering at the University of Michigan and the University of Michigan's Transportation Research Institute. He received a BS in Electrical Engineering and Computer Sciences and an Honors Degree in Physics in May 2006, an MS degree in Electrical Engineering in May 2009, and a PhD in Electrical Engineering in December 2012 all from the University of California, Berkeley before completing a postdoc in the Computer Science and Artificial Intelligence Laboratory at the Massachusetts Institute of Technology. His research interests include dynamical systems, optimization, and robotics especially with applications involving human interaction.