TY - JOUR
T1 - Iterative learning control with parameter estimation for non-repetitive time-varying systems
AU - Wang, Lei
AU - Huangfu, Ziwei
AU - Li, Ruiwen
AU - Wen, Xiewen
AU - Sun, Yuan
AU - Chen, Yiyang
N1 - Funding Information:
This work was supported by the National Science Foundation of China (No. 62103293 ), the Natural Science Foundation of Jiangsu Province in China (No. BK20210709 ), the Wuxi Innovation and Entrepreneurship Fund “Taihu Light” Science and Technology (Fundamental Research) Project, China under Grant (No. K20221045 ), the Start-up Fund for Introducing Talent of Wuxi University, China (No. 2021r045 ), and the Innovative Leading Talents in Universities of Xishan Talents Program, China (No. 2022xsyc001 ).
Publisher Copyright:
© 2024 The Franklin Institute
PY - 2024/2
Y1 - 2024/2
N2 - This paper presents an extension method of iterative learning control (ILC) to address the applications associated with non-repetitive time-varying systems (NTVSs). Conventional ILC approaches employ fixed nominal system models, but non-repetitive time-varying models may lead to accumulated model uncertainties, which fails to satisfy the robust convergence conditions. To tackle this issue, a novel ILC algorithm with parameter estimation is proposed using back propagation neural network. This algorithm incorporates an approach that utilizes Bayesian regularization training mechanism to accurately estimate non-repetitive time-varying parameters. Through comprehensive experiment on Monolithic XY Stage, the performance of proposed algorithm is validated to demonstrate its feasibility and effectiveness while handling tasks on NTVSs.
AB - This paper presents an extension method of iterative learning control (ILC) to address the applications associated with non-repetitive time-varying systems (NTVSs). Conventional ILC approaches employ fixed nominal system models, but non-repetitive time-varying models may lead to accumulated model uncertainties, which fails to satisfy the robust convergence conditions. To tackle this issue, a novel ILC algorithm with parameter estimation is proposed using back propagation neural network. This algorithm incorporates an approach that utilizes Bayesian regularization training mechanism to accurately estimate non-repetitive time-varying parameters. Through comprehensive experiment on Monolithic XY Stage, the performance of proposed algorithm is validated to demonstrate its feasibility and effectiveness while handling tasks on NTVSs.
KW - Bayesian regularization
KW - Iterative learning control
KW - Neural network
KW - Non-repetitive time-varying systems
KW - Parameter estimation
UR - http://www.scopus.com/inward/record.url?scp=85182726746&partnerID=8YFLogxK
U2 - 10.1016/j.jfranklin.2024.01.011
DO - 10.1016/j.jfranklin.2024.01.011
M3 - Journal article
AN - SCOPUS:85182726746
SN - 0016-0032
VL - 361
SP - 1455
EP - 1466
JO - Journal of the Franklin Institute
JF - Journal of the Franklin Institute
IS - 3
ER -