Abstract
Certainty equivalence adaptive controllers are analysed using a “data-driven Riccati equation”, corresponding to the model-free Bellman equation used in Q-learning. The equation depends quadratically on data correlation matrices. This makes it possible to derive simple sufficient conditions for stability and robustness to unmodeled dynamics in adaptive systems. The paper is concluded by short remarks on how the bounds can be used to quantify the interplay between excitation levels and robustness to unmodeled dynamics.
| Original language | English |
|---|---|
| Title of host publication | Proceedings of Machine Learning Research |
| Editors/authors | Alessandro Abate, Mark Cannon, Kostas Margellos, Antonis Papachristodoulou |
| Publisher | ML Research Press |
| Pages | 504-513 |
| Number of pages | 10 |
| Volume | 242 |
| Publication status | Published - 2024 |
| Event | 6th Annual Learning for Dynamics and Control Conference, L4DC 2024 - Oxford, United Kingdom Duration: 2024 Jul 15 → 2024 Jul 17 |
Conference
| Conference | 6th Annual Learning for Dynamics and Control Conference, L4DC 2024 |
|---|---|
| Country/Territory | United Kingdom |
| City | Oxford |
| Period | 2024/07/15 → 2024/07/17 |
Subject classification (UKÄ)
- Computational Mathematics
Free keywords
- adaptive control
- dual control
- linear quadratic control
- online learning
Fingerprint
Dive into the research topics of 'A Data-driven Riccati Equation'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver