System identification for achieving robust performance

Selecting Uncertainty Structures in Identification for Robust Control with an Automotive Application

“…Furthermore, the minimization of (7) over P, subject to (2), is at the heart of iterative identification and robust control approaches, including de Callafon and Van den Hof [1997], Oomen and Bosgra [2012]. An important advantage of the criterion (7) is that the robust control design (3) and identification problem (7) may be solved alternately, leading to a monotonously converging iterative procedure Bayard et al [1992].…”

“…Specifically, a main result of Oomen and Bosgra [2012] reveals that 16th IFAC Symposium on System Identification Brussels, Belgium. July 11-13, 2012…”

“…In Oomen and Bosgra [2012], a new model uncertainty structure has been presented that has distinct advantages from a robust performance perspective. A key ingredient of this uncertainty structure is a new coprime factorization that arises in a novel connection between controlrelevant identification of nominal models and coprime factor identification, extending and providing new insights in earlier results, including Schrama [1992].…”

“…A second ingredient is a certain (W u ,W y )-normalized RCF, see Oomen and Bosgra [2012]. Specifically, the pair {N…”

“…These coprime factor-based uncertainty structures have been further refined towards dual-Youla uncertainty structures in, e.g., Anderson [1998], Douma and Van den Hof [2005], that improve the connection between identification and control by explicitly considering the closed-loop operation of the system. Recently, in Oomen and Bosgra [2012], these coprime factor-based uncertainty structures are further refined to explicitly connect the size of uncertainty and the control criterion. An essential advantage of the latter structure is that it provides an inherent scaling of the uncertainty channels that is essential for the nonconservative identification of model sets.…”