Learning Attack Generation
Attack generation for complex nonlinear system
Image credit: CPS
Motivation
- Model-based attack generation approach requires knowledge of system’s model, but for some complex system, high-fiedilty model is hard to be obtained;
- For nonlinear systems, attack generation problem is a highly nonconvex problem
- No prior labelled attack dataset
Approach
Applications
Improve attack detection precision
-
Attack generator performance:
-
Train attack detector
Explore vulnerability space of networked systems
- Power Grid
- Gas pipline
[under development] - Connected Vehicle
[under development]
Relevant paper
- Y. Zheng, Ali Sayghe, OM Anubi, “Algorithm Design for Resilient Cyber-Physical Systems using an Automated Attack Generative Model”, Engineering Applications of Artificial Intelligence. [Under review], (2023)
- Y. Zheng, OM Anubi, “Data-driven Vulnerability Analysis of Networked Pipeline System”, IEEE Conference on Control Technology and Applications (CCTA). [Under review], (2023).
- Y. Zheng, S. Vedula, OM Anubi, Learning to Attack Nonlinear Networked Cyber-Physical Systems, IEEE Transactions on Control System Technology. [Under review], (2023)
