Passivity Analysis And A Security Design Of Cyber Physical System

Security design for cyber physical systems under attack is mainly considered in this paper. First, the condition to preserve dissipativity and stability under approximation is analyzed. Based on prior knowledge of dissipativity parameters for the approximation system, nonlinear stable model predictive control strategy proposed in this paper guarantees the stability and dissipativity for the real dynamic cyber physical system. This scheme can deal with nonlinear, non-passive systems with disturbances. Not only can this scheme guarantee the robustness for the real large-scale networked system, but also break the limitation that good performance of traditional model predictive control strategy rely tightly on the accuracy of the model. This provides an insight for us to deal with complex cyber physical systems when we cannot get the accurate model for them.Besides, based on passivity theorems, event-triggering scheme proposed in this paper preserves dissipativity under discretization. The triggering condition proposed guarantees that the closed-loop system is still QSR-dissipative.Thirdly, we propose a resilient mechanism to monitor and defense expected or unexpected attacks and describe the transitions between each mode though a hybrid automaton. Based on traditional disturbance observer, the attack monitor designed can detect the existence of exogenous worst-case attacks and track their trajectories. Based on H-infinity control theorem, the defense mechanism designed obtains the optimal strategies for attacker and defender for systems under attack. The solutions obtained are also Nash equilibrium in this zero-sum game. Finally the resilient mechanism is implemented in a cyber-physical vehicle example. The objective of this paper is to come up with a secure, hybrid, intelligent design such that system defenses the unexpected attacks and preserves the desired performance under attack. This topic combines modeling, information theory, optimal control, game theory, which is a cross-frontier disciplines for control, information and system engineering researchers.