filtering for uncertain stochastic systems subject to sensor nonlinearities

This work considers the filtering problem for uncertain stochastic systems subject to sensor nonlinearities. It may be seen from simulation results in this work that the traditional filtering method based on linear measurement may not provide a reliable solution to this problem due to the existence of the nonlinear characteristic of sensors. In the system under consideration, there exist time-varying parameter uncertainties, and state and external-disturbance-dependent noise. Robust filters are constructed for both continuous and discrete stochastic systems, such that the resultant filtering error systems are robustly stochastically stable with a prescribed H∞-disturbance attenuation performance. Finally, some simulation results with deterministic or stochastic disturbance signals are given to illustrate the proposed method.