Multiple Model Filtering in the Presence of Gaussian Mixture Measurement Noises

A simplified multiple model filter is developed for discrete-time systems in the presence of Gaussian mixture measurement noises. Theoretical analysis proves that the proposed filter has the same estimation performance as the interacting multiple model filter at the price of less computational cost. Numerically robust implementation of the filter is presented to meet practical applications. An example on bearings-only guidance demonstrates the effect of the proposed algorithm.