Distributed Actuator Fault Detection of LDTV Systems Using Relative Output Measurements
Ping Wang, Chengpu Yu · 2021
This paper investigates the distributed actuator fault detection problem for linear discrete time-varying heterogeneous multi-agent systems using relative output measurements. Firstly, an augmented model is built for each agent by stacking all local relative output measurements for detection scheme design. Secondly, to deal with the affect caused by the absence of fault signals in the observation equation, a distributed fault detection residual generator is constructed by introducing the future output measurements and a feedback term. In addition, an indefinite quadratic minimum problem is proposed using the finite horizon H∞filtering framework, for which the necessary and sufficient minimum conditions are derived using the Krein-space theory. Then, a computationally efficient recursive algorithm is developed to determine the residual at each time step. Afterwards, a distributed fault detection scheme is provided to alert the fault occurrence. Finally, a simulation example is given to show the effectiveness of the proposed algorithm.