A State-Space Approach to H∞ Actuator Fault Detection for Sampled-Data Systems

Actuator fault detection problem for sampled-data systems is investigated in an H-infinity setting, where the effect from exogenous disturbance and uncertainties in the initial state to the residual is attenuated in a prescribed level. Such a problem is first formulated as a modified sampled-data control problem and it turns out the fault detector owns the jump system's structure. One sufficient existence condition and the design algorithm for the detector are then both provided in terms of solution to a set of Riccati differential equation with finite discrete jumps. The proposed approach for sampled-data actuator fault detection problem is a simple state-space method and it improves the existed results in two folds. First, robustness to the measurement noise and uncertainties in the initial state is considered. Second, parameters of the sampled-data system are allowed to be time-varying.