A robust high-order recursive quadratic algorithm for linear-in-the- parameters models

Abstract

In this paper a robust k-order Recursive Quadratic algorithm for a large class of linear-in-the-parametes models is introduced to enhance Recursive Quadratic algorithm's robustness. The features of the algorithm are discussed and the convergence of the algorithm has been proven in this paper. Several examples are included to demonstrate the efficiency by comparing the result with the conventional least square algorithm and the effectiveness of the robust Recursive Quadratic algorithm.