Identification of general time-varying systems

This thesis investigates the identification of general time-variant systems using Hilbert-Huang transform (HHT) and Bayesian inference. An HHT-based method is extended for general time variant systems. It is then combined with a Bayesian inference based method to obtain the distributions of system parameters. Another Bayesian inference based method is proposed using intrinsic mode functions to formulate the likelihood function. Numerical simulation results demonstrate the proposed methods.