Constrained state estimation using the Unscented Kalman Filter
Rambabu Kandepu, Lars Imsland, Bjarne A. Foss · 2008
A simple procedure to include state inequality constraints in the unscented Kalman filter is proposed. With this procedure, the information of active state constraints influences the state covariance matrix, resulting in better estimates. In a numerical example, the approach outperforms the extended Kalman filter implemented with constraint handling via ldquoclippingrdquo.