Design and Comparative Study of Limited Complexity Bounding Error Identification Algorithms

Abstract Bounding error algorithms consist on updating the exact membership set or one of its classical approximations . The exact membership set updating algorithm suffers from the computing amount which disables its application in real time control and the algorithms updating the approximations may yield a conservative domain. This paper presents algorithms that update the polytope by limiting its complexity. Three methods proposed in the literature are presented and a new one is proposed. Compared to classical algorithms, these techniques provide better results. Numerical as well as practical simulations are presented to illustrate the efficiency and the comparative study of such proposed algorithms.