Fault detection for networked systems with random packet losses, sensor saturation and multiplicative noises

This paper is concerned with the H∞ fault detection (FD) for a class of networked systems with random packet losses, sensor saturation and multiplicative noises. The network with both output measurement and control packet losses is modeled with two independent Bernouli distributed white sequences. H∞ filtering theory is adapted to formulate the problem of fault detection for networked systems. The output measurement is affected by sensor saturation which is described by sector-nonlinearities, and the multiplicative noises are modeled as a form of Gaussian white noise. The purpose of the addressed problem is to design a fault detection filter such that, the fault detection dynamic system is exponentially stable in the mean square, and the error between the residual value and fault value is made as small as possible. A sufficient condition for FD filter is derived by solving the linear matrix inequality (LMI). Finally, a numerical example is illustrated to show the effectiveness of the designed method.