Disturbance state distinction algorithm and its application to estimation of time delay with inertia error

This paper proposes a novel algorithm to detect information of time delay and inertia errors of the nominal model in the disturbance observer. The proposed method adopts the hidden Markov model to relate the states causing disturbances with the observed disturbances. Output distribution of the hidden disturbance state can be defined based on the relationship between the observed disturbance and time delay/inertia errors. The proposed algorithm is compared with the Interacting Multiple Model algorithm. Simulation results verify effectiveness and properties of the proposed algorithm.