Informative Conditions for Identification of MISO ARMAX Model in Closed-loop Systems

In system identification, a data set needs to be informative to ensure that the identification criterion has a unique global minimum asymptotically, and the parameter estimation is consistent. In this work, meaningful sufficient and necessary informative conditions for a data set in multi-input single-output (MISO) closed-loop systems are obtained for the identification of ARMAX model. Compared with the results in the existing literature, our theoretical results provide precisely quantitative conditions on the complexity of controllers to generate informative experiment without using external excitation. It is shown that the newly-proposed theoretical results are in agreement with the simulation results. (C) 2018, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.