Robust adaptive identification of slowly time-varying parameters with bounded disturbances

In this paper, the robustness limitations of current recursive identification algorithms are highlighted. Then, it is shown how a particular nonrecursive identification algorithm dramatically improves the robustness to bounded disturbances, noise and slowly time-varying parameters. only at the expense of performing an on-line eigenvalue decomposition on a symmetric semidefinite positive matrix. Furthermore, this algorithm do not require any a priori knowledge of a bound on the disturbance and noise and of a bound on the parameter values.