A Sub-Domain-Awareness Adaptive Probabilistic Event-Triggered Policy for Attack-Compensated Output Control of Markov Jump CPSs With Dynamically Matching Modes
Haiyang Chen,Guangdeng Zong,Xiang Liu,Xudong Zhao,Ben Niu,Fangzheng Gao
DOI: https://doi.org/10.1109/tase.2023.3296259
IF: 6.636
2023-01-01
IEEE Transactions on Automation Science and Engineering
Abstract:This paper investigates the attack-compensated output control problem for a class of Markov jump cyber-physical systems (MJCPSs) subject to mismatched modes. A sub-domain-awareness adaptive probabilistic event-triggered mechanism (APETM) is innovatively developed to fully enhance the control performance of the networked control system. Here, sub-domain-awareness means the network’s communication ability, which may be affected by time delay and network resource. To defend against the cyber-attacks imposed on the communication channel, a predictor-based compensator is constructed to mitigate the impact of attacks on the control performance. Different from the existing works, the mismatch degree between the controller and the system, governed by a random variable obeying any discrete-time distribution over $[0, 1]$ , is changing dynamically. A mismatched output feedback controller is designed based on the APETM and attack-compensator, and an augmented closed-loop system is obtained. Both time delay-and triggering threshold-dependent Lyapunov functionals are introduced to perform the asymptotical stability analysis. With a separation technique, a solving algorithm is provided which converts the non-convex analysis conditions into traceable ones. Finally, simulations are conducted to validate the proposed results through a mass-spring-damping system model. Note to Practitioners—Cyber-physical systems (CPSs) have found many applications in the fields such as smart grid, gas distribution systems, and automated manufacturing systems. The open communication network deployed in CPSs face two main challenges including the security of data transmission and the efficiency of resource utilization. Owing to the practical environment and other factors, system parameters may change abruptly, which can be well defined by a Markov model. In the resultant Markov jump system, it is typically assumed that the mode mismatch degree between the controller and the system remains the same, which is unrealistic in a time-varying network environment. Motivated by these observations, this paper investigates the secure output control of Markov jump CPSs (MJCPSs) subject to mismatched modes and unknown cyber-attacks. An adaptive probabilistic event-triggered mechanism is innovatively proposed, which compared with the existing mechanisms strikes a higher level balance between the communication burden and the control performance. And attack-compensated mismatched output controllers are developed to stabilize MJCPSs and meanwhile preserve the security level. Finally, a mass-spring-damping system is adopted to verify the application effectiveness of the proposed results. Note that communication traffic control and secure control are finely involved based on CPSs, making our proposed results more applicable to practical systems.
automation & control systems