Combined state and parameter estimation for not fully observable dynamic systems. (September 2020)

• Record Type: Journal Article
• Authors: Backi, Christoph Josef; Gravdahl, Jan Tommy; Skogestad, Sigurd
• Is Part Of: IFAC journal of systems and control. Volume 13(2020)
• Publication Date: 2020-09
• Subjects: Parameter identification -- Estimation -- Kalman filter -- System identification; Automatic control -- Periodicals; Relay control systems -- Periodicals; Embedded computer systems -- Periodicals; Feedback control systems -- Periodicals; Artificial intelligence -- Periodicals; Artificial intelligence; Automatic control; Embedded computer systems; Feedback control systems; Relay control systems; Electronic journals; Periodicals; 629.89
• DOI: 10.1016/j.ifacsc.2020.100103
• Languages: English
• ISSNs: 2468-6018

Abstract

Abstract: In this paper, a simple, yet novel method for state estimation and parameter identification for dynamic systems is presented. Apart from providing estimates of non-measurable state variables, the algorithm is also capable of estimating (constant) system parameters. The estimation algorithm is split in two parts. Firstly, an extended Kalman filter, whose state-space-model is augmented with quasi-linear expressions for parameter values, providing estimates for the state variables and the augmented parameter values. Secondly, a Monte-Carlo-fashioned approach, which identifies the rest of the parameter values that were not included in the augmentation of the state-space model. The Monte-Carlo-approach minimizes an objective function (the error between the measured and the estimated state variable). It is shown that the algorithm is capable of estimating the state- and parameter-values in a satisfying manner. The method is best applied offline and the theoretical developments will be demonstrated in case studies.