Extreme-learning-machine-based robust integral terminal sliding mode control of bicycle robot. (April 2022)

• Record Type: Journal Article
• Authors: Chen, Long; Yan, Bin; Wang, Hai; Shao, Ke; Kurniawan, Edi; Wang, Guangyi
• Is Part Of: Control engineering practice. Volume 121(2022)
• Publication Date: 2022-04
• Subjects: Balancing -- Integral terminal sliding mode (ITSM) -- Bicycle robot (BR) -- Extreme learning machine (ELM); Automatic control -- Periodicals; 629.89
• DOI: 10.1016/j.conengprac.2022.105064
• Languages: English
• ISSNs: 0967-0661

Abstract

Abstract: In this paper, an extreme-learning-machine (ELM)-based robust integral terminal sliding mode (ITSM) control scheme is developed for a bicycle robot (BR) to achieve balancing target. First, the bicycle robot equipped with a reaction wheel is formulated by a second-order mathematical model with uncertainties. Then, an ITSM controller is designed for the balancing control of the BR, where an ELM scheme is designed as a compensator for estimating lumped uncertainties of the system. The stability proof of the closed-loop control system is presented based on Lyapunov theory. Comparative experimental results are demonstrated to verify the superior balancing performance of the proposed control. Highlights: Newly formulated robot model is composed of the nominal model and parameter uncertainty. Integral terminal sliding mode (ITSM) feedback control input ensures finite-time convergence. Extreme learning machine (ELM) is designed as a compensator to estimate the lumped uncertainty. Closed-loop stability proof of ELM-based ITSM control is rigorously given via Lyapunov theory. Comparative real-time experiments show the excellent performance of the ELM controller.