Iterative filter with finite measurements for suddenly maneuvering targets

The finite impulse response (FIR) structure is extended to Markov linear systems (JMLSs) based on the optimal unbiased FIR filters developed for a general time-variant state-space system and the IMM algorithm. The proposed algorithm degenerates to the IMM approach as the estimation horizon length approaches infinity. Researchers introduce the IMM filter briefly and present the problem ormulation. The researchers demonstrate that it is known that the optimal Bayesian estimation for JMLSs in the minimum mean-square error sense is practical intractable because its computational complexity grows exponentially with the time step.