Sensor Fault Estimation for Nonlinear Complex Networks With Time Delays Under Saturated Innovations: A Binary Encoding Scheme

In this paper, the fault estimation problem is dealt with for a class of nonlinear time-delayed complex networks with bounded stochastic noises and sensor faults under the binary encoding scheme (BES). A first-order differential model with a random bias is utilized to describe the sensor fault to cater for real practice. By using the BES, the measurement signal is first encoded to a bit string and then transmitted to the estimator where the bit string is decoded. During the transmission, random bit errors may occur in the bit string due to the channel noises. A set of Bernoulli distributed random variables is employed to characterize the occurrence of the bit errors. To enhance the robustness of the estimation performance to the random bit errors, the saturation function is introduced in the estimator to constrain the innovations. The target of this paper is to design a fault estimator such that the estimation error dynamics is exponentially ultimately bounded in mean square with a minimized ultimate bound. Sufficient conditions are first developed for the existence of the desired fault estimator via stochastic analysis and inequality manipulations. Then, the estimator gain matrices are obtained by solving a set of matrix inequalities. Finally, a simulation example is presented to illustrate the effectiveness of the proposed fault estimation method.