Auto-updating of sampling time redesign for system identification under parameter uncertainty

Wang, Ke and Yue, Hong; (2019) Auto-updating of sampling time redesign for system identification under parameter uncertainty. In: 2019 25th International Conference on Automation and Computing (ICAC). IEEE, GBR, pp. 169-174. ISBN 9781861376657 (https://doi.org/10.23919/IConAC.2019.8895129)

[thumbnail of Wang-Yue-ICAC2019-Auto-updating-of-sampling-time-redesign-for-system-identification]
Preview
Text. Filename: Wang_Yue_ICAC2019_Auto_updating_of_sampling_time_redesign_for_system_identification.pdf
Accepted Author Manuscript

Download (493kB)| Preview

Abstract

In this paper, the mid-term experimental redesign of sampling strategy has been studied to improve design robustness under model uncertainties. With this method, the whole design time horizon is divided into several zones. Parameter estimation is made with the data in each zone and the updated model is used for experimental design in the next zone until the experiment is complete. A novel auto-updating strategy is developed to determine the length of each zone to assure the identifiability for parameter estimation. No pre-settings from users are required as in previous redesign algorithms. Simulation studies on an enzyme reaction system have been conducted. The results demonstrate that, compared to the conventional offline design and the standard online redesign methods, the proposed mid-term redesign with auto-updating of zones produces data that lead to more accurate parameter identification under large uncertainties.