A recursive interpolatory algorithm for ℓ1 robust identification

This paper proposes an on-line recursive interpolatory algorithm for the ℓ¹ robust identification problem. The new idea is to describe the experimental information and a prior information about system and noise by geometric approach. Then, we can determine whether or not the new experiment data contains innovations by the containing relation of model set determined by old data and the set formed from new data, meanwhile, the uncertainty of model set is reduced by effectively using the innovations contained in new data. Thus, the identification algorithm proposed in this paper can effectively use the experimental data and improve the identification precision. Finally analysis of the proposed recursive algorithm in this paper is carried out and a simulation example is presented to demonstrate the effectiveness of the algorithm.