Integrated navigation algorithm based on IMM-UKF

A new unscented Kalman filter (UKF) based on interacting multiple model (IMM) is presented to solve the problem of nonlinear filtering and noise modeling. The uncertainty of the noise can be described by a set of switching models. In every model a UKF is running, and the UKF for nonlinear filtering can achieve accuracy at least to the second order. The output of the IMM-UKF is the weighted sum of a bank of parallel unscented Kalman filters. The self-adaptive filtering for different noises can be performed by the adjustment of all models weights. The application of the algorithm on integrated navigation system shows a high precision and switching speed, so it is applicable to dynamic systems.