Event-triggered fault estimation for nonlinear systems with missing measurements

Abstract

In this paper, the joint fault and state estimation problem is investigated for a class of nonlinear systems with event-triggered transmissions and missing measurements. In the proposed event-triggered transmission scheme, in order to reduce unnecessary network traffic, the current measurement is released only when it changes greatly from the previously transmitted one. A Bernoulli distributed sequence taking values on 0 or 1 is introduced to govern possible missing measurements in the transmission. Special effort is made to obtain and then minimize certain upper bound of the estimation error covariance in the simultaneous presence of the linearization errors and imperfect measurement transmissions. It is noticeable that, in the proposed method, the traditional assumption on the availability on the probability density functions of the states and the innovations conditional on the measurements is no longer needed, and therefore the application scope is much widened. Moreover, the fault and states can be jointly estimated, thereby providing a way of simultaneously monitoring the system and diagnosing the faults. The estimator gain is calculated via solving two recursive matrix equations, and the corresponding algorithm is therefore suitable for online applications. An illustrative example is provided to show the effectiveness of the proposed algorithm.