A System Level Approach to Distributed and Adaptive Control

Speaker: Nikolai Matni
Affiliation: UC Berkeley

Abstract: Modern cyber-physical systems such as the smart-grid, automated transportation systems and the internet of things are of huge scale, safety-critical, operate in uncertain and ever-changing environments, and have control logic that is often sparse, distributed, uncertain and limited.  Despite these challenges, system designers must be able to provide strong guarantees of performance, robustness, and safety, as failures can be catastrophic.  This talk presents progress towards developing a unified theoretical framework for the design of large-scale data-driven cyber-physical systems that leverages our recently developed System Level Approach (SLA) to Controller Synthesis.  The SLA provides a transparent connection between system structure, constraints, and uncertainty and their effects on controller synthesis, implementation, and performance.  This transparent connection has allowed us to make significant breakthroughs in distributed optimal control and adaptive control.  In the first part of this talk, we show how the SLA allows for localized optimal controllers to be synthesized using convex programming, thus extending the performance and robustness guarantees of optimal/robust control, under mild and practically relevant assumptions, to systems of arbitrary size.  We illustrate the usefulness of this approach with a frequency regulation problem in the power-grid, and show how it can be used to systematically explore tradeoffs in controller performance, robustness, and synthesis/implementation complexity. The second part of this talk presents a modern take on robust and adaptive control that merges techniques from statistical learning theory and robust/optimal control to derive performance guarantees for a “Learning Linear Quadratic Regulator” (LQR) optimal control problem wherein the system model to be controlled is unknown. Leveraging the SLA, we establish, to the best of our knowledge, the first end-to-end baselines for learning in an LQR problem that do not require restrictive or unrealistic assumptions.  We conclude with our vision for a contemporary theory of distributed adaptive control, and outline ongoing efforts in extending the previous results to incorporate the safety and performance guarantees of other control paradigms, such as model predictive control.

Biography:  Nikolai is a postdoctoral scholar in EECS at UC Berkeley working with Benjamin Recht. Prior to that, he was a postdoctoral scholar in Computing and Mathematical Sciences at the California Institute of Technology. He received the B.A.Sc. and M.A.Sc. in Electrical Engineering from the University of British Columbia, and the Ph.D. in Control and Dynamical Systems from the California Institute of Technology in June 2016 under the advisement of John C. Doyle. His research interests broadly encompass the use of learning, layering, dynamics, control and optimization in the design and analysis of large-scale data-driven cyber-physical systems. He was awarded the IEEE CDC 2013 Best Student Paper Award, the IEEE ACC 2017 Best Student Paper Award (as co-advisor), and was an Everhart Lecture Series speaker at Caltech.

For more information, contact Prof. Paulo Tabuada (tabuada@ucla.edu)

Date/Time:
Date(s) - Jan 30, 2018
11:00 am - 12:30 pm

Location:
EE-IV Shannon Room #54-134
420 Westwood Plaza - 5th Flr., Los Angeles CA 90095