Non-fragile Joint State and Fault Estimation for Complex Networks with Missing Measurements and Switching Topology

In this paper, a non-fragile joint state and fault estimation problem is investigated for time-varying complex networks with switching topology and randomly occurring nonlinearities under the missing measurements. The phenomena of the switching topology, randomly occurring nonlinearities and missing measurements are described by three mutually independent Bernoulli random variables, where the uncertain occurrence probabilities are depicted. In addition, the parameter perturbations of the estimator gain matrix are characterized by a set of zero-mean multiplicative noises. A novel time-varying estimation method is designed and an upper bound of estimation error covariance matrix is obtained by solving two recursive matrix equations. Further, the desired estimator gain matrix is parameterized by minimizing the trace of the obtained upper bound. Subsequently, a sufficient condition is given to ensure the boundedness of the upper bound matrix by using the mathematical induction method. Lastly, a simulation study is given to verify the feasibility of the proposed estimation scheme.