Identification methods for time-delay systems based on the redundant rules

摘要

Two identification methods based on the redundant rules are proposed for a class of systems with unknown time-delays. One is the recursive least squares based redundant parameter method and the other is the variational Bayesian based redundant model method. These two methods can estimate the unknown parameters and the unknown time-delay simultaneously. Furthermore, a comparative study reveals that the variational Bayesian based redundant model method is more effective than the recursive least squares based redundant parameter method. The simulation made further guarantees the usefulness of the proposed methods.