Attacking the V: On the Resiliency of Adaptive-Horizon MPC
Inspired by the emerging problem of CPS security, we introduce the concept of controller-attacker games. A controller-attacker game is a two-player stochastic game, where the two players, a controller and an attacker, have antagonistic objectives. A controller-attacker game is formulated in terms of a Markov Decision Process (MDP), with the controller and the attacker jointly determining the MDP’s transition probabilities. We also introduce the class of controller-attacker games we call V-formation games, where the goal of the controller is to maneuver the plant (a simple model of flocking dynamics) into a V-formation, and the goal of the attacker is to prevent the controller from doing so. Controllers in V-formation games utilize a new formulation of model-predictive control we have developed called Adaptive-Horizon MPC (AMPC), giving them extraordinary power: we prove that under certain controllability conditions, an AMPC controller can attain V-formation with probability 1. We evaluate AMPC’s performance on V-formation games using statistical model checking. Our experiments demonstrate that (a) as we increase the power of the attacker, the AMPC controller adapts by suitably increasing its horizon, and thus demonstrates resiliency to a variety of attacks; and (b) an intelligent attacker can significantly outperform its naive counterpart.
Research supported in part by the Doctoral Program Logical Methods in Computer Science and the Austrian National Research Network RiSE/SHiNE (S11412-N23) project funded by the Austrian Science Fund (FWF) project W1255-N23, AFOSR Grant FA9550-14-1-0261 and NSF Grants CCF-1423296, CNS-1423298, IIS-1447549, CNS-1446832, CNS-1445770, CNS-1445770.
- 2.Checkoway, S., McCoy, D., Kantor, B., Anderson, D., Shacham, H., Savage, S., Koscher, K., an, A.C., Roesner, F., Kohno, T.: Comprehensive experimental analyses of automotive attack surfaces. In: USENIX Security (2011)Google Scholar
- 3.Condliffe, J.: A 100-drone swarm, dropped from jets, plans its own moves, MIT Technology Review, January 2017Google Scholar
- 4.Davidson, D., Wu, H., Jellinek, R., Ristenpart, T., Singh, V.: Controlling UAVs with sensor input spoofing attacks. In: Proceedings of 10th USENIX Workshop on Offensive Technologies, WOOT 2016, Austin, TX, August 2016Google Scholar
- 5.Droge, G., Egerstedt, M.: Adaptive time horizon optimization in model predictive control. In: 2011 American Control Conference (ACC), pp. 1843–1848. IEEE (2011)Google Scholar
- 8.Kennedy, J., Eberhart, R.: Particle swarm optimization. In: Proceedings of 1995 IEEE International Conference on Neural Networks, pp. 1942–1948 (1995)Google Scholar
- 9.Krener, A.J.: Adaptive horizon model predictive control, arXiv preprint (2016). arXiv:1602.08619
- 11.Narendra, K.S.: Adaptive control using neural networks. In: Neural Networks for Control, pp. 115–142. MIT Press (1990)Google Scholar
- 12.Negash, L., Kim, S.H., Choi, H.L.: An unknown-input-observer based approach for cyber attack detection in formation flying UAVs. In: AIAA Infotech (2016)Google Scholar
- 13.Ogata, K.: Modern Control Engineering: Instrumentation and Controls Series. Prentice Hall, Upper Saddle River (2010)Google Scholar
- 14.Pajic, M., Weimer, J., Bezzo, N., Tabuada, P., Sokolsky, O., Lee, I., Pappas, G.J.: Robustness of attack-resilient state estimators. In: 5th ACM/IEEE International Conference on Cyber-Physical Systems (ICCPS) (2014)Google Scholar
- 15.Park, J., Ivanov, R., Weimer, J., Pajic, M., Lee, I.: Sensor attack detection in the presence of transient faults. In: 6th ACM/IEEE International Conference on Cyber-Physical Systems (ICCPS) (2015)Google Scholar
- 19.Yang, J., Grosu, R., Smolka, S.A., Tiwari, A.: Love thy neighbor: V-formation as a problem of model predictive control. In: LIPIcs-Leibniz International Proceedings in Informatics, vol. 59. Schloss Dagstuhl-Leibniz-Zentrum fuer Informatik (2016)Google Scholar
- 20.Yang, J., Grosu, R., Smolka, S.A., Tiwari, A.: V-formation as optimal control. In: Proceedings of the Biological Distributed Algorithms Workshop (2016)Google Scholar