State estimation for a kind of non-uniform sampling dynamic system

This article is concerned with the state estimation problem for a kind of non-uniform sampling dynamic system. The system is described at a certain sampling rate in discrete form. A sensor observes the target successively, while the samples are obtained non-uniformly and the sampling points in each sampling periods may be different. To generate the optimal state estimate, state prediction and innovation are carried out step-by-step similar to Kalman filter (KF), but before innovation in each sampling period, measurements should be properly disposed and augmented. It is shown that our main results improve and extend the existing KF for which the samples are obtained multirate non-uniformly. Measurements randomly missing with Bernoulli distribution are also considered in this article. Finally, the feasibility and efficiency of the presented algorithm is illustrated by a numerical simulation example.