Event-triggered hybrid stubborn ESO for networked systems with uncertain disturbances and stochastic deception attacks.

Abstract

In this paper, the state estimation problem is investigated for a general class of nonlinear networked systems subject to both external disturbances and stochastic deception attacks. In the presence of deception attacks, a novel hybrid stubborn extended state observer (ESO) is established to estimate the states and total disturbances, simultaneously. In addition, the event-triggered mechanism (ETM) is introduced utilizing the estimation errors to relieve the burden of the transmission networks. Then, an inter-trigger output predictor is adopted based on the estimation state of the hybrid stubborn ESO to predict the information between two consecutive triggering moments. To overcome the disadvantage of the outliers induced by the deception attack, the saturation nonlinearity is adopted as the gain function of the hybrid stubborn ESO. Sufficient conditions are established to ensure that the estimation error dynamics is locally mean-square bounded in a domain of attraction, and then an iterative linear matrix inequality (ILMI) approach is employed to design the desired hybrid stubborn ESO. Moreover, the Zeno behavior can be avoided when the designed ETM is utilized. Some numerical simulations are conducted to demonstrate the validity of the proposed methodology.