State estimation for jump Markov linear systems with uncompensated biases

This paper studies the problem of state estimation for jump Markov linear systems with uncompensated biases. By describing the state and the measurement biases as additive random variables, a suboptimal filter has been developed by applying the basic interacting multiple model (IMM) approach. To derive a precise representation of the biases contributions to the state estimation, three auxiliary matrices are introduced with respect to the correlation between the state estimation errors and the biases, which helps to derive mode-conditioned estimates in the framework of the IMM. A numerical example involving tracking a maneuvering target is provided to compare the performance of the proposed filter with that of the augmented state filter.