Fault detection for networked control systems based on dynamic feedback observer

For the networked control system with both Markov random delay and packet dropout, a fault detection method based on dynamic feedback observer is studied. The networked control system is modeled as a Markov jump system. With the output error as the input signal, the dynamic feedback observer generates the dynamic residual signal and feed it back to the equations of state estimation and output estimation. The problem of fault detection filter design is transformed to the H ∞ filtering problem. The solvable condition of the fault detection filter and further the solutions are derived in terms of LMI. A numerical example illustrates the effectiveness of the proposed approach.