Filtering of nonlinear systems with measurement loss by RUKF-IMM

From interacting multiple model (IMM) estimation, the states in nonlinear systems with measurement loss were estimated by IMM approach. The model set in the IMM approach contained two models, one of them corresponded to the situations with measurement loss and the other one corresponded to the situations without measurement loss. Final estimates were obtained based on the fusion of the two model estimates. This approach improved the stability of the estimator in systems with missed measurements. Each filter in IMM used the randomized unscented Kalman (RUKF) to estimate the states. RUKF eliminated the system errors caused by the unscented Kalman fitler (UKF). Thus the accuracy of the estimation was also improved. The simulation results show the proposed approach is more stable and accurate than the traditional one-model mixture estimation approaches in nonlinear systems with measurement loss.