26 January 2018 in IEEE Transactions on Automatic Control
In this paper, a security problem in networked control systems (NCS) is studied. In a standard linear quadratic Gaussian (LQG) control scenario in NCS, a so-called false data injection attack could be launched by a malicious attacker to deteriorate the system performance without being detected. To defend against such attacks, a defender on the NCS side needs to allocate defense resources among the sensors to secure the data, and the defense investment determines the costs of compromising certain sensors. After observing the defender’ action, the attacker decides the target sensors to compromise. While both sides are subject to the resource constraints, the interactive decision-making between the defender and the attacker is investigated in a Stackelberg game (leader–follower game) framework. The optimal solutions for both sides under different types of budget constraints are analyzed. Simulation examples are provided to illustrate the main results.
1 January 2014 in Automatica
1 June 2014 in Automatica
19 September 2017 in IEEE Transactions on Automatic Control
In this note, we consider event-triggered extended state observer (ET-ESO) design for a continuous time nonlinear system with uncertainty and disturbance. Since the convergence of the ET-ESO depends on the design of the triggering condition directly, the goal is to design a realizable triggering condition not relying on the state of the plant. For the extended state observer considered, an event-triggered transmission strategy is proposed such that the observation error is uniformly bounded. Furthermore, under mild conditions, the observation error can be guaranteed to be bounded at any time. It is shown that there is no Zeno behavior for the event-based transmission strategy. The obtained theoretical results are validated through numerical simulations.
12 November 2014 in IEEE Transactions on Automatic Control
Motivated by challenges in state estimation with event-based measurement updates, the properties of the exact and approximate set-valued Kalman filters with multiple sensor measurements for linear time-invariant systems are investigated in this work. First, we show that the exact and the proposed approximate set-valued filters are independent of the fusion sequence at each time instant. Second, the boundedness of the size of the set of estimation means is proved for the exact set-valued filter. For the approximate set-valued filter, if the closed-loop matrix is contractive, then the set of estimation means has a bounded size asymptotically; otherwise a nonsingular linear transform is constructed such that the size of the set of estimation means for the transformed states is asymptotically bounded. Third, the effect of set-valued measurements on the size of the set of estimation means is analyzed and conditions for performance improvement in terms of smaller size of the set of estimation means are proposed. Finally, the results are applied to event-based estimation, which allow the event-triggering conditions to be designed by considering requirements on performance and communication rates. The efficiency of the proposed results are illustrated by simulation examples and comparison with the approximate event-based MMSE estimator and the Kalman filter with intermittent observations.
15 January 2018 in IEEE Transactions on Control of Network Systems
In this paper, we study the worst-case consequence of innovation-based integrity attacks with side information in a remote state estimation scenario where a sensor transmits its measurement to a remote estimator equipped with a false-data detector. If a malicious attacker is not only able to compromise the transmitted data packet but also able to measure the system state itself, the attack strategy can be designed based on the intercepted data, the sensing data, or alternatively the combined information. Surprisingly, we show that launching attacks using the combined information are not always optimal. First, we characterize the stealthiness constraints for different types of attack strategies to avoid being noticed by the false-data detector. Then, we derive the evolution of the remote estimation error covariance in the presence of attacks, based on which the worst-case attack policies are obtained by solving convex optimization problems. Furthermore, the closed-form expressions of the worst-case attacks are obtained for scalar systems and the attack consequences are compared with the existing work to determine which strategy is more critical in deteriorating system performance. Simulation examples are provided to illustrate the analytical results.