I’m interested in the very broad area of complex interconnected systems, where (potentially heterogeneous) agents interact influencing each other. I’m fascinated by how control theoretical tools can help framing networked systems in a mathematically sound structure, and the insights therein obtained.
Control of Self-Driving Cars
During my time at nuTonomy, I worked mostly on trajectory tracking control for self-driving cars. Specifically, I was interested in obtaining a more human-like and comfortable driving experience. I achieved these goals by means of numerical optimization techniques, as well as a direct usage of the perception pipeline output. Some of the results have been presented at ITSC 2019 (check out the Publications section).
Scalable Rebalancing Techniques for AMoD Systems
If not properly managed, a fleet of self-driving cars becomes imbalanced (that is, cars end up where no one needs them, and customers wait in vain somewhere else). In the literature there are several approaches on how to “rebalance” Autonomous Mobility on Demand (AMoD) systems, however they always assume a specific network structure. During my Master’s Thesis, I focused on using optimal control techniques to improve the quality of the rebalancing performed, and at the same time to allow for general network structures, obtaining a more scalable algorithm. I was supervised by Prof. Emilio Frazzoli, Prof. Melanie N. Zeilinger, Dr. Andrea Carron and Claudio Ruch, all affiliated with the Institute for Dynamic Systems and Control (IDSC), ETH Zürich.