Fault detection in finite frequency domain for networked control systems with missing measurements

Abstract

This paper is concerned on the fault detection (FD) problem in finite frequency domain for networked control systems (NCSs) with missing measurements. By virtue of the stopping time, the considered NCSs are firstly modeled as Markov jumping systems (MJSs). The notion of finite frequency stochastic H− index is then introduced to measure the sensitivity of the residuals. Taking into account a new sensor fault model, which is valid to express the failures of stuck, loss of effectiveness as well as outage ones, a novel FD scheme is developed with simultaneous consideration of sensitivity performance and attenuation performance in finite frequency domain, such that it is valid for all admissible sensor faults. In addition, new convex conditions in terms of linear matrix inequalities (LMIs), which can be reduced to some previous results, are presented to cope with this FD problem. Further, fault detection filters (FDFs) can be constructed by solving the derived LMIs. Finally, such an FD scheme is utilized to an aircraft model, and the effectiveness of proposed method is demonstrated by the simulation results.