Stability Analysis of Networked Control Systems under Denial of Service Attacks using Switching System Theory

With the development of computer networks, packet-based data transmission has found its way to Cyber-Physical Systems (CPS) and especially, networked control systems (NCS). NCSs are distributed industrial processes in which sensors and actuators exchange information between the physical plant and the controller via a network. Any loss of data or packet in the network links affects the performance of the physical system and its stability. This loss could be due to natural congestions in network or a result of intentional Denial of Service (DoS) attacks. In this paper, we analytically study the stability of NCSs with the possibility of data loss in the feed-forward link by modelling the system as a switching one. When data are lost (or replaced with a jammed or bogus invalid signal/packet) in the forward link, the physical system will not receive the control input sent from the controller. In this study, NCS is regarded as a stochastic switching system by using a two-position Markov jump model. In State 1, the control signal/packet passes through and gets to the system, while in State 2, the signal or packet is lost. We analyze the stability of system in State 2 by considering the situation as an open-loop control scenario with zero input. The proposed stochastic switching system is studied in both continuous and discrete-time spaces to see under what conditions it satisfies Lyapunov stability. The stability conditions are obtained according to random dwell times of the system in each state. Finally, the model is simulated on a DC motor as the plant. The results confirm the correctness of the obtained stability conditions.