Optimization And Security Research For Networked Control Systems Via A Delta Operator Approach

Networked control systems(NCSs) bring many advantages such as lower expenses,higher maintainability, higher interactivity and higher ability of fault diagnosis than the traditional point-to-point control systems. The research on NCSs has been one of the most important aspects in the area of control. However, there are many problems and challenges in NCSs. Insertion of communication networks in the feedback-loop system breaks the integrity of the control systems and causes a lot of challenging issues, such as time delays,packet-dropout, packet out of order, communication limitation and sampling, transmitting methods. Exposing the control systems to public networks can increase the mobility and inter-operability of the control systems. The control systems are not isolated systems anymore by transmitting the control and measurement signals over public networks. The probability of being attacked has increased greatly. The design and analysis of NCSs under cyber attacks is a new cross field. Most of the researches are just beginning and many problems still need to be addressed.Controllers are analyzed and designed for NCSs in this paper with random time delays, packet-dropout and disturbances. The NCSs are modeled by delta operator approach, and controllers are designed for NCSs with the game theoretic approach to achieve the optimal performance of each subsystem. Defense strategy and attack strategy are presented to guarantee NCSs operating safely when the NCSs are under Do S attacks.Our main results are given as follows:Firstly, a multi-tasking framework is proposed for NCSs with external disturbance.The delta operator approach is employed to realize nonuniform sampling according to dynamic network environment for the multi-tasking NCSs. A ?-Nash equilibrium(NE) is proposed as the objective for dealing with external disturbance in the NCSs. Control strategies for the multi-tasking NCSs are developed via a game theoretic approach. An upper bound for the ?-NE is also provided analytically to guarantee robustness of the NCSs. Some simulation results on two-area load frequency control(LFC) are given to show the validity and advantage of the proposed method.Secondly, NCSs with multiple disturbances and unknown non-linearity are investigated in this part. A delta operator approach is employed to construct the system model to overcome numerical stiffness at high frequency sampling. An observer is proposed to estimate the disturbance generated by exogenous systems. An optimal controller is developed to optimize system performance by integrating feed forward compensation of disturbance observer. A ?-level linear quadratic optimal control strategy is presented to characterize the influence of observation error, unmatched disturbance and unknown non-linearity. An upper bound for the ?-level is provided analytically to guarantee the robustness of NCSs. Simulation results on LFC are given to show the validity and advantage of the main results in this part.Thirdly, we consider the security problem of wireless NCSs. Comparing with the wired systems, wireless systems are more unreliale, the research on security of wireless networked systems is vary important. Attackers attack the transmitter in NCSs through launching jamming signal. Establish the system model of the cyber-layer, and both the goals of the transmitter and the attackers are maximizing their own utility. The communication quality of cyber-layer will influence the packets’ transmission of the control-layer. With the coupled relationship, we present a sufficient condition to guarantee the stability of the NCSs with actuator saturation. Simulation results are given to illustrate the effectiveness of the proposed method.Finally, we consider the problem of resilient control of NCSs under Do S attacks via a unified game approach. The Do S attacks lead to extra constraints in NCSs, where the packets may be jammed by a malicious adversary. Considering the attack induced packet-drop, optimal control strategies with multi-tasking and central-tasking structures are developed using game theory in the delta domain, respectively. Based on the optimal control structures, we propose optimality criteria and algorithms for cyber defenders and Do S attackers. We apply this method to temperature regulation of Heating, Ventilation and Air Conditioning(HVAC) system and ball and beam system to illustrate the solution methodology.