A new convergent recursive identification algorithm
Tien Chang Hsia · 1983
Presented in this paper is a least-squares based recursive algorithm for system parameter identification free of the usual positive-real conditions required by many similar recursive algorithms. The algorithm is globally asymptotic convergent and robust. Simulation studies have also demonstrated that the new algorithm has a faster converging rate than similar algorithms. Comparison results with the RELS and Landau's algorithms are presented.