Least squares estimation for a class of non-uniformly sampled systems based on the hierarchical identification principle

Abstract

This paper presents a novel hierarchical least squares algorithm for a class of non-uniformly sampled systems. Based on the hierarchical identification principle, the identification model with a high dimensional parameter vector is decomposed into a group of submodels with lower dimensional parameter vectors. By using the least squares method to identify the submodels and taking a coordinated measure to address the associated items between the submodels, all the system parameters can be estimated. The proposed algorithm can save the computation cost. The performance analysis indicates that parameter estimates converge to their true values. The simulation tests confirm the convergence results.