State estimation for discrete time-delayed non-fragile switched neural network with parameter uncertainties and sojourn probabilities
K. Sai Uma Maheswari, Ramalatha Marimuthu, S. N. Shivapriya · AIP conference proceedings · 2020
This paper designs to investigate the non-fragility nature of a state estimator with parameter uncertainties and sojourn probabilities for a class of discrete-time neural networks. The uncertainties formed by the norm boundedness on the parameter, consists of the nonlinear and the linear parts in the state estimator of the neuron with the error considered by the variation in the estimator gain. A new switched system is so modeled by sojourn probability method for all admissible parameter uncertainties and gain variations which guarantees the resulting error system is asymptotic stable in the mean square. By applying Lyapunov function technique, the desired designing of the state estimator with the gain matrix's characterization is done explicitly. Finally, a numerical simulation is illustrated for the developed state estimation approach.