Regret-Optimal Defense Against Stealthy Adversaries: A System Level Approach
Authors:
Hiroyasu Tsukamoto,
Joudi Hajar,
Soon-Jo Chung,
Fred Y. Hadaegh
Abstract:
Modern control designs in robotics, aerospace, and cyber-physical systems rely heavily on real-world data obtained through system outputs. However, these outputs can be compromised by system faults and malicious attacks, distorting critical system information needed for secure and reliable operation. In this paper, we introduce a novel regret-optimal control framework for designing controllers tha…
▽ More
Modern control designs in robotics, aerospace, and cyber-physical systems rely heavily on real-world data obtained through system outputs. However, these outputs can be compromised by system faults and malicious attacks, distorting critical system information needed for secure and reliable operation. In this paper, we introduce a novel regret-optimal control framework for designing controllers that make a linear system robust against stealthy attacks, including both sensor and actuator attacks. Specifically, we present (a) a convex optimization-based system metric to quantify the regret under the worst-case stealthy attack (the difference between actual performance and optimal performance with hindsight of the attack), which adapts and improves upon the $\mathcal{H}_2$ and $\mathcal{H}_{\infty}$ norms in the presence of stealthy adversaries, (b) an optimization problem for minimizing the regret of (a) in system-level parameterization, enabling localized and distributed implementation in large-scale systems, and (c) a rank-constrained optimization problem equivalent to the optimization of (b), which can be solved using convex rank minimization methods. We also present numerical simulations that demonstrate the effectiveness of our proposed framework.
△ Less
Submitted 14 September, 2024; v1 submitted 25 July, 2024;
originally announced July 2024.
Robust Optimal Network Topology Switching for Zero Dynamics Attacks
Authors:
Hiroyasu Tsukamoto,
Joshua D. Ibrahim,
Joudi Hajar,
James Ragan,
Soon-Jo Chung,
Fred Y. Hadaegh
Abstract:
The intrinsic, sampling, and enforced zero dynamics attacks (ZDAs) are among the most detrimental stealthy attacks in robotics, aerospace, and cyber-physical systems. They exploit internal dynamics, discretization, redundancy/asynchronous actuation and sensing, to construct disruptive attacks that are completely stealthy in the measurement. They work even when the systems are both controllable and…
▽ More
The intrinsic, sampling, and enforced zero dynamics attacks (ZDAs) are among the most detrimental stealthy attacks in robotics, aerospace, and cyber-physical systems. They exploit internal dynamics, discretization, redundancy/asynchronous actuation and sensing, to construct disruptive attacks that are completely stealthy in the measurement. They work even when the systems are both controllable and observable. This paper presents a novel framework to robustly and optimally detect and mitigate ZDAs for networked linear control systems. We utilize controllability, observability, robustness, and sensitivity metrics written explicitly in terms of the system topology, thereby proposing a robust and optimal switching topology formulation for resilient ZDA detection and mitigation. Our main contribution is the reformulation of this problem into an equivalent rank-constrained optimization problem (i.e., optimization with a convex objective function subject to convex constraints and rank constraints), which can be solved using convex rank minimization approaches. The effectiveness of our method is demonstrated using networked double integrators subject to ZDAs.
△ Less
Submitted 14 September, 2024; v1 submitted 25 July, 2024;
originally announced July 2024.
