Event-triggered fault detection for linear stochastic systems

Abstract

In this paper, the event-triggered fault-detection problem is investigated for Itô-type stochastic systems. A filter structure is used to construct a residual model for fault detection. For limited network resources, a novel event-triggered strategy is proposed, while a trigger condition is utilized to determine and transmit the useful data through the network. Compared with the traditional periodic release strategy, the proposed design approach can significantly reduce the utilization of network bandwidth. A new condition is proposed to achieve the mean-square asymptotical stability of the residual model with the desired fault-detection objective. A novel algorithm is derived to obtain the parameters of the filter and the event detector. Two simulation cases are provided to illustrate the effectiveness of the design strategy.