Distributed fault detection for discrete-time nonlinear systems: An innovation-based approach

This paper addresses the problem of fault detection for a class of discrete-time nonlinear systems when using multiple sensors. A parallel distributed architecture is used to derive the state estimates, in which the unscented Kalman filter (UKF) is employed to deal with the nonlinear filtering problem. By augmenting the normalized innovation sequences, which can be derived in the UKF, into an innovation matrix, the statistical properties of this innovation matrix are used to develop fault detection rules. A numerical example is provided to verify the effectiveness of the proposed method.