Abstract
This talk provides an overview of our recent results on controlling uncertain systems with guaranteed constraint satisfaction. It is by now well understood how to address additive and parametric uncertainties in a model predictive control (MPC) scheme to provide robust constraint satisfaction guarantees. However, in the presence of general dynamic uncertainties, like modelling errors or uncertain time delays, these guarantees are no longer valid. Therefore, we develop a new robust MPC framework based on integral quadratic constraints (IQCs) accounting for all uncertainties that can be described by an IQC, which includes dynamic, time-varying, and even nonlinear uncertainties. In order to guarantee pointwise-in-time bounds, like constraints, we extend the IQC framework to the analysis of energy-to-peak and peak-to-peak gains and provide a controller design procedure to minimize these performance measures. Building on these results we further develop an output-feedback tube-based MPC scheme that is robustly stabilizing, recursively feasible and ensures robust constraint satisfaction.
Biographical Information
Lukas Schwenkel is a Ph.D. candidate at the Institute for Systems Theory and Automatic Control, University of Stuttgart, advised by Prof. Frank Allgöwer. He is also a member of the International Max Planck Research School for Intelligent Systems (IMPRS-IS). He received his M.Sc. in Engineering Cybernetics from the University of Stuttgart in 2019. His main research interests include robust control and model predictive control.