Design of unscented Kalman filter with correlative noises

An Unscented-Kalman Filter(UKF) for a class of nonlinear discrete-time systems with correlated noises is designed to deal with the problem of nonlinear filtering failure in conventional UKF when system noise is correlated with measurement noise. Recursive filtering equations of UKF with correlated noises are given based on the minimum meansquare- error estimation; and unscented transformation(UT) is applied to the calculation the posterior means and covariances of the nonlinear system states. The proposed UKF breaks through the limitations on the conventional UKF that the system noise and measurement noise must be uncorrelated Gauss white noises, thus extending the applications of the conventional UKF. A simulation example shows its feasibility and effectiveness.