Robust fault detection for networked control systems with nonlinear disturbances and imperfect measurements

Abstract

In this article, the robust fault detection (FD) problem is investigated for networked control systems with nonlinear disturbances and imperfect measurements. Under the consideration of packet-dropout compensation, a new measurement model is proposed to take the time-varying delay, random packet dropout and quantisation effect into account simultaneously. After properly augmenting the states of the original system and the FD filter, the robust FD problem is formulated as an auxiliary H _∞ filtering problem for a stochastic parameter system with time-varying delays and uncertainties. A sufficient condition for the existence of the robust FD filter is derived in terms of linear matrix inequalities, which depends on both the network status and the quantisation density. A numerical example is provided to illustrate the effectiveness of the proposed method.

Keywords:

• fault detection
• nonlinear networked control system
• signal quantisation
• time-varying delay
• packet-dropout compensation
• linear matrix inequality

Acknowledgements

This work was supported by the National Natural Science Foundation of China under Grants 60874053 and 61034006.