Event triggered fault detection in linear systems using Generalized likelihood ratio test

摘要

An event triggered fault detection problem for linear systems is addressed using generalized likelihood ratio test. Using measurements obtained from any general event sampling mechanism and assuming known fault magnitude, fault occurrence time is estimated. While estimating the fault occurrence time, state estimates are required in computing likelihood functions. For measurements satisfying the event triggering mechanism, states are estimated with the obtained measurements, while for the other measurements, event triggering information is used to obtain the state estimates. The effectiveness of the proposed approach is demonstrated using an illustrative example.