Combined cubature Kalman and smooth variable structure filtering: A robust nonlinear estimation strategy

In this paper, nonlinear state estimation problems with modeling uncertainties are considered. As demonstrated recently in literature, the cubature Kalman filter (CKF) provides the closest known approximation to the Bayesian filter in the sense of preserving second-order information contained in noisy measurements under the Gaussian assumption. The smooth variable structure filter (SVSF) has also been recently introduced and has been shown to be robust to modeling uncertainties. In an effort to utilize the accuracy of the CKF and the robustness of the SVSF, the CKF and SVSF have been combined resulting in an algorithm referred to as the CK–SVSF. The robustness and accuracy of the CK–SVSF was validated by testing it on two different computer problems, namely, a target tracking problem and the estimation of the effective bulk modulus in an electrohydrostatic actuator. •A new nonlinear estimation strategy is presented based on combining elements of the cubature Kalman filter and the smooth variable structure filter.•The filters are combined based on the time‐varying smoothing boundary layer concept.•Two nonlinear estimation problems are studied and the results demonstrated that the proposed filter (CK–SVSF) yields improved results over other popular strategies.