Multiagent decision making: learning from observations
Automatic Control Laboratory
Swiss Federal Institute of Technology Zurich
We began using robots to help us with simple and repetitive tasks. Now, we are introducing automation in ever increasing safety critical and complex tasks: intelligent transportation networks, smart power grid, robotics search and rescue, personalized medicine. How do we ensure that these autonomous systems contribute to making our societies safer and more efficient? To answer this question, we have to address two main control challenges: 1) The autonomous systems need to complete complex tasks in partially known changing environments; 2) Network of interacting autonomous systems need to coordinate their decision making and optimise their choices using only local information. Guided by addressing these two challenges, I will present the two main threads of my research. Specifically, I discuss my latest works on data-driven safe control synthesis in uncertain environments and on game theory and mechanism design for coordinating multi-agent decision making. In both cases, I present the underlying theoretical tools and discuss potential applications.
Maryam Kamgarpour is an assistant professor at ETH Zurich, Automatic Control Laboratory. She obtained Doctor of Philosophy in Engineering from the University of California, Berkeley (2011) and Bachelor of Applied Sciences from the University of Waterloo (2005). She addresses multiagent decision making and safe control synthesis using game theory, mechanism design, online stochastic and robust optimization. Her work is applied to electrical power systems, transportation systems, and rescue robotics. She is the recipient of NASA High Potential Individual Award, NASA Excellence in Publication Award (2010) and the European Union (ERC) Starting Grant 2015.