TR2020-092
Learning-based Parameter-Adaptive Reference Governors
-
- "Learning-based Parameter-Adaptive Reference Governors", American Control Conference (ACC), DOI: 10.23919/ACC45564.2020.9147615, July 2020, pp. 956-961.BibTeX TR2020-092 PDF
- @inproceedings{Chakrabarty2020jul,
- author = {Chakrabarty, Ankush and Berntorp, Karl and Di Cairano, Stefano},
- title = {Learning-based Parameter-Adaptive Reference Governors},
- booktitle = {American Control Conference (ACC)},
- year = 2020,
- pages = {956--961},
- month = jul,
- publisher = {IEEE},
- doi = {10.23919/ACC45564.2020.9147615},
- url = {https://www.merl.com/publications/TR2020-092}
- }
,
- "Learning-based Parameter-Adaptive Reference Governors", American Control Conference (ACC), DOI: 10.23919/ACC45564.2020.9147615, July 2020, pp. 956-961.
-
MERL Contacts:
-
Research Areas:
Abstract:
Reference governors (RGs) provide an effective method for ensuring safety via constraint enforcement in closedloop control systems. When the parameters of the underlying systems are unknown, but constant or slowly-varying, robust formulations of RGs that consider only the worst-case effect may be overly conservative and exhibit poor performance. This paper proposes a parameter-adaptive reference governor (PARG) architecture that is capable of generating safe trajectories in spite of parameter uncertainties without being as conservative as robust RGs. The proposed approach leverages on-line data to inform algorithms for robust parameter estimation. Subsequently, confidence bounds around parameter estimates are fed to supervised machine learners for approximating robust constraint admissible sets leveraged by the PARG. While initially, due to the absence of on-line data, the PARG may be as conservative as a robust RG, as more data is gathered and the confidence bounds become tighter, such conservativeness reduces, as demonstrated in a simulation example.
Related News & Events
-
NEWS Yebin Wang delivered an invited industry talk at the 1st IEEE Industrial Electronics Society Annual On-Line Conference Date: December 9, 2022 - December 11, 2022
MERL Contact: Yebin Wang
Research Areas: Communications, Control, OptimizationBrief- Future factory, in the era of industry 4.0, is characterized by autonomy, digital twin, and mass customization. This talk, titled "Future factory automation and cyber-physical system: an industrial perspective," focuses on tackling the challenges arising from mass customization, for example reconfigurable machine controller and material flow.
-
NEWS Ankush Chakrabarty gave an invited talk at University of Illinois at Chicago Date: April 9, 2021
MERL Contact: Ankush Chakrabarty
Research Areas: Control, Machine Learning, Multi-Physical Modeling, OptimizationBrief- Ankush Chakrabarty, a Research Scientist at MERL's Multiphysical Systems (MS) Team, gave an invited talk on "Learning for Control and Estimation using Digital Twins" at the Department of Electrical and Computer Engineering Seminar Series organized at UIC. The talk proposed new learning-based control/estimation architectures that can utilize simulation data obtained from digital twins to add self-optimization and constraint-enforcement features to grey/black-box control systems.
-
NEWS MERL researchers presented 10 papers at American Control Conference (ACC) Date: July 1, 2020 - July 3, 2020
Where: Denver, Colorado (virtual)
MERL Contacts: Mouhacine Benosman; Ankush Chakrabarty; Stefano Di Cairano; Yebin Wang; Avishai Weiss
Research Areas: Control, Machine Learning, OptimizationBrief- At the American Control Conference, MERL presented 10 papers on subjects including autonomous-vehicle decision making and motion planning, nonlinear estimation for thermal-fluid models and GNSS positioning, learning-based reference governors and reference governors for railway vehicles, and fail-safe rendezvous control.