Resiliency of Perception-Based Controllers Against Attacks

Amir Khazraei, Henry Pfister, Miroslav Pajic
Proceedings of The 4th Annual Learning for Dynamics and Control Conference, PMLR 168:713-725, 2022.

Abstract

This work focuses on resiliency of learning-enabled perception-based controllers for nonlinear dynamical systems. We consider systems equipped with an end-to-end controller, mapping the perception (e.g., camera images) and sensor measurements to control inputs, as well as a statistical or learning-based anomaly detector (AD). We define a general notion of attack stealthiness and find conditions for which there exists a sequence of stealthy attacks on perception and sensor measurements that forces the system into unsafe operation without being detected, for any employed AD. Specifically, we show that systems with unstable physical plants and exponentially stable closed-loop dynamics are vulnerable to such stealthy attacks. Finally, we use our results on a case-study.

Cite this Paper

BibTeX
@InProceedings{pmlr-v168-khazraei22a, title = {Resiliency of Perception-Based Controllers Against Attacks}, author = {Khazraei, Amir and Pfister, Henry and Pajic, Miroslav}, booktitle = {Proceedings of The 4th Annual Learning for Dynamics and Control Conference}, pages = {713--725}, year = {2022}, editor = {Firoozi, Roya and Mehr, Negar and Yel, Esen and Antonova, Rika and Bohg, Jeannette and Schwager, Mac and Kochenderfer, Mykel}, volume = {168}, series = {Proceedings of Machine Learning Research}, month = {23--24 Jun}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v168/khazraei22a/khazraei22a.pdf}, url = {https://proceedings.mlr.press/v168/khazraei22a.html}, abstract = {This work focuses on resiliency of learning-enabled perception-based controllers for nonlinear dynamical systems. We consider systems equipped with an end-to-end controller, mapping the perception (e.g., camera images) and sensor measurements to control inputs, as well as a statistical or learning-based anomaly detector (AD). We define a general notion of attack stealthiness and find conditions for which there exists a sequence of stealthy attacks on perception and sensor measurements that forces the system into unsafe operation without being detected, for any employed AD. Specifically, we show that systems with unstable physical plants and exponentially stable closed-loop dynamics are vulnerable to such stealthy attacks. Finally, we use our results on a case-study.} }
Endnote
%0 Conference Paper %T Resiliency of Perception-Based Controllers Against Attacks %A Amir Khazraei %A Henry Pfister %A Miroslav Pajic %B Proceedings of The 4th Annual Learning for Dynamics and Control Conference %C Proceedings of Machine Learning Research %D 2022 %E Roya Firoozi %E Negar Mehr %E Esen Yel %E Rika Antonova %E Jeannette Bohg %E Mac Schwager %E Mykel Kochenderfer %F pmlr-v168-khazraei22a %I PMLR %P 713--725 %U https://proceedings.mlr.press/v168/khazraei22a.html %V 168 %X This work focuses on resiliency of learning-enabled perception-based controllers for nonlinear dynamical systems. We consider systems equipped with an end-to-end controller, mapping the perception (e.g., camera images) and sensor measurements to control inputs, as well as a statistical or learning-based anomaly detector (AD). We define a general notion of attack stealthiness and find conditions for which there exists a sequence of stealthy attacks on perception and sensor measurements that forces the system into unsafe operation without being detected, for any employed AD. Specifically, we show that systems with unstable physical plants and exponentially stable closed-loop dynamics are vulnerable to such stealthy attacks. Finally, we use our results on a case-study.
APA
Khazraei, A., Pfister, H. & Pajic, M.. (2022). Resiliency of Perception-Based Controllers Against Attacks. Proceedings of The 4th Annual Learning for Dynamics and Control Conference, in Proceedings of Machine Learning Research 168:713-725 Available from https://proceedings.mlr.press/v168/khazraei22a.html.

Related Material