An output delay approach to sensor fault detection for non-uniformly sampled-data systems
基于输出时滞方法的非均匀采样数据系统传感器故障检测

In this paper, the problem of sensor fault detection in general non-uniformly sampled-data systems is investigated. First, the output delay approach is used to model non-uniformly sampled-data systems as continuous ones with time-varying delay output. Then, based on the input-output approach, a fault detection filter is designed to guarantee not only the effects of continuous-time process noise and discrete-time measurement noise on residual to satisfy a prescribed H-infinity performance, but also the l-two gain from sensor fault to residual to be greater than a prescribed value. The existence condition of such a fault detection filter is given in terms of matrix inequalities(MIs). Furthermore, an iterative algorithm which transform MIs into the solvable linear MIs(LMIs) is proposed to make a tradeoff between the noise robustness and fault sensitivity. Finally, simulation results of a certain type of flight control systems are presented to show the validity of the proposed method.