Recursive algorithms for identification in closed loop: A unified approach and evaluation (Q1372747)

The authors provide a unified framework for plant model identification. After stating the basic equations associated with the identification system, they present two families of algorithms: one is the family of the closed-loop output-error recursive identification algorithms, and the other one is the family of filtering open-loop recursive identification algorithms. Moreover, for the validation of a plant model in closed loop they state the statistical validation tests that include the uncorrelation test and the whiteness test, the pole closeness validation test, and the time-domain validation test. Finally, they present the comparative evaluation of the various algorithms that include the closed-loop output-error (CLOE) algorithm, the filtered CLOE algorithm, the adaptive filtered CLOE algorithm, the extended CLOE algorithm, the generalized CLOE algorithm, and the filtered open-loop identification algorithm via three simulated examples and on a real system which is a very flexible transmission system.