Robust convergence of two-stage nonlinear algorithms for identification in H∞
Two-stage nonlinear algorithms for the problem of identification in H∞ are analyzed in terms of the Fourier transform of the window function. Conditions characterizing the robust convergence of the two-stage nonlinear algorithm are derived. The sufficiency of the conditions holds for a general class of window functions while the necessity is proved for robust convergence of the first stage of the algorithm under the mild additional restriction that the window function is even symmetric. Some improved upper bounds for the worst case identification error are also obtained. © 1992.
Publication Source (Journal or Book title)
Systems and Control Letters
Gu, G., Khargonekar, P., & Li, Y. (1992). Robust convergence of two-stage nonlinear algorithms for identification in H∞. Systems and Control Letters, 18 (4), 253-263. https://doi.org/10.1016/0167-6911(92)90055-W