Sensor Fault Reconstruction via Event-Triggered Sliding Mode Observer

This paper presents the sensor fault reconstruction of the linear time-invariant systems under resource constraint scenarios. The resource constraint issues, such as communication burden and computational burden, are addressed here by adopting an event-triggering technique. Under this aperiodic data transmission strategy, we propose the sliding mode observer, which is designed based on the dynamics of an augmented system. This proposed observer estimates the plant states practically and enables the reconstruction of the sensor fault signal with the desired accuracy. Furthermore, to address the problem of higher magnitude chattering in state estimation, we introduce a strategy for switching the observer discontinuous gain to a smaller value based on the convergence time of the state estimation error to its ultimate bound. Finally, a simulation study is provided to demonstrate the proposed sensor fault reconstruction technique.