Abstract
Data-driven predictive control (DPC) is a feedback control method for systems with unknown dynamics. It repeatedly optimizes a system's future trajectories based on past input-output data. We develop a numerical method that computes poisoning attacks that inject additive perturbations to the online output data to change the trajectories optimized by DPC. This method is based on implicitly differentiating the solution map of the trajectory optimization in DPC. We demonstrate that the resulting attacks can cause an output tracking error one order of magnitude higher than random perturbations in numerical experiments.
| Original language | English (US) |
|---|---|
| Title of host publication | 2023 American Control Conference, ACC 2023 |
| Publisher | Institute of Electrical and Electronics Engineers Inc. |
| Pages | 545-550 |
| Number of pages | 6 |
| ISBN (Electronic) | 9798350328066 |
| DOIs | |
| State | Published - 2023 |
| Externally published | Yes |
| Event | 2023 American Control Conference, ACC 2023 - San Diego, United States Duration: May 31 2023 → Jun 2 2023 |
Publication series
| Name | Proceedings of the American Control Conference |
|---|---|
| Volume | 2023-May |
| ISSN (Print) | 0743-1619 |
Conference
| Conference | 2023 American Control Conference, ACC 2023 |
|---|---|
| Country/Territory | United States |
| City | San Diego |
| Period | 5/31/23 → 6/2/23 |
Bibliographical note
Publisher Copyright:© 2023 American Automatic Control Council.