A robust predictor for dead-time systems based on the Kalman filter. Issue 25 (2018)

• Record Type: Journal Article
• Authors: Lima, Bruno M.; Lima, Daniel M.; Normey-Rico, Julio E.
• Is Part Of: IFAC-PapersOnLine. Volume 51:Issue 25(2018)
• Page Start: 24
• Page End: 29
• Publication Date: 2018
• Subjects: delay compensation -- observers -- robust control -- Kalman filters; Automatic control -- Periodicals; 629.805
• DOI: 10.1016/j.ifacol.2018.11.076
• Languages: English
• ISSNs: 2405-8963

Abstract

Abstract: Dead time processes are common in the industry and represent a challenge for control. One way to avoid (or attenuate) the problem is to use a predictor structure alongside the controller. In this paper, an observer-predictor structure based on a steady-state Kalman filter is explored. It is shown that this structure is equivalent to the Filtered Smith Predictor (FSP) for linear systems, hence, all the tools used to analyze the closed-loop properties of the FSP can also be used in the proposed predictor, specially the frequency domain ones. With the results of this analysis, tuning guidelines for the predictor are given to achieve different closed-loop characteristics, e.g., better disturbance rejection or better robustness. Furthermore, the proposed predictor can easily be used with all kind of systems, including unstable and non-minimum phase systems.