Resilient Minimum Entropy Filter Design for Non-Gaussian Stochastic Systems

In this paper, the resilient minimum entropy filter problem is investigated for the stochastic systems with non-Gaussian disturbances. The goal of designing the filter is to guarantee that the entropy of the estimation error is monotonically decreasing, moreover, the error system is exponentially ultimately bounded in the mean square. Based on the entropy performance function, a filter gain updating algorithm is presented to make the entropy decrease at every sampling instant k. Then the boundedness of the gain updating law is analyzed using the kernel density estimation technique. Furthermore, a suboptimal resilient filter gain is designed in terms of LMI. Finally, a simulation example is given to show the effectiveness of the proposed results.