An analaysis of recursive parametric estimation methods for large-scale time-verying systems

Abstract EN

In this paper, three recursive parametric estimation methods are analysed to solve the estimate problem for a class of large-scale time-varying systems, which can be decomposed into several interconnected single-input single-output systems.

These methods, which are based upon the recursive least squares techniques, incorporate weighted factors into the adaptation gain matrices of their recursive parametric estimation algorithms.

It is about the Recursive Least Squares method with Forgetting Factor (RLSFF), the Recursive Least Squares method with Weighted Parameters (RLSWF) and the Recursive Least Squares method using Kalman Filter (RLSKF).

Recursive parametric estimation algorithms are consequently developed that permit to follow the parametric variations of each interconnected system in order to ensure a good quality of the estimate.

The techniques of the practical implementation of the developed parametric estimation algorithms are given.

Furthermore, simulated results are presented illustrating the performance and the effectiveness of these algorithms.