Extended Kalman filter and Takagi-Sugeno fuzzy observer for a strip winding system. (1st December 2022)

• Record Type: Journal Article
• Authors: Szedlak-Stinean, Alexandra-Iulia; Precup, Radu-Emil; Petriu, Emil M.; Roman, Raul-Cristian; Hedrea, Elena-Lorena; Bojan-Dragos, Claudia-Adina
• Is Part Of: Expert systems with applications. Volume 208(2022)
• Publication Date: 2022-12-01
• Subjects: Extended Kalman filters -- Low-cost automation -- Mechatronics -- Nonlinear observers -- Strip winding systems -- Takagi-Sugeno fuzzy observers -- Variable parameters; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2022.118215
• Languages: English
• ISSNs: 0957-4174

Abstract

Graphical abstract: Zoomed plots of angular velocity versus time. Highlights: A detailed mathematical model of a Strip Winding System (SWS) is derived. The bump-less switching between three control algorithms is ensured. An extended Kalman filter approach for the SWS is designed. A Takagi-Sugeno Fuzzy Observer (TSFO) for the SWS is designed. Linear matrix inequalities are involved in the design of the TSFO. Abstract: This paper proposes two nonlinear estimation approaches, namely based on Extended Kalman Filter (EKF) and a Takagi-Sugeno Fuzzy Observer with 32 rules (TSFO-32), for a Strip Winding System (SWS) characterized by variable reference input, variable moment of inertia with constant increasing tendency and variable parameters. The SWS is a complex and nonlinear mechatronic system viewed as a controlled process, which wraps a strip with constant linear velocity on a reel, and the variable radius modifies both the angular velocity and the moment of inertia. The motivation of using EKF is the zero stationary error at low speeds. The motivation of using TSFO-32 is that once the Takagi-Sugeno fuzzy models are obtained, various analysis and design tools initially developed for linear systems, which facilitate the observation and/or synthesis of the controller for complex nonlinear systems, can be adapted appropriately and used in these nonlinear systems. Therefore, these tools simplify the design as well-established approaches and algorithms are available. The fuzzy control system stability and observer design conditions are derived and expressed as linear matrix inequalities. The efficiency of TSFO-32 is discussed in this paper in terms of setting a certain convergence rate. The performance of the two nonlinear estimation approaches is validated by means of digital simulations conducted for three values of the moment of inertia.