Filter-based recursive Bayesian algorithm with modified covariance resetting for non-uniformly sampled data systems

Abstract

To identify a system with non-uniformly sampled data, a recursive Bayesian algorithm combined dynamic filter with covariance resetting is proposed. First, the input-output data is filtered by the estimated noise transfer function, and the system is decomposed into two fictitious sub-systems with a low dimension. Second, the estimated variance of the noise is employed in the proposed algorithm to improve the estimates. Furthermore, an efficient covariance resetting strategy is integrated into the algorithm to increase the convergence rate. Finally, the proposed algorithm is validated by a numeric example.