Online Parameter Estimation of Linear Systems Using A New Designed Extended State Observer

In this paper, an observer-based online parameter estimation method for linear systems is proposed. It is shown that when the sampling interval is small enough, the sign of unmeasurable state estimation errors can be estimated and used to construct a new extended sliding mode observer with finite ti...

Full description

Saved in:
Bibliographic Details
Main Authors: Zhang, Dongsheng, Man, Zhihong, Wang, Hai, Zheng, Jinchuan, Cao, Zhenwei, Wang, Song
Format: Conference Proceeding
Language:English
Subjects:
Convergence
extended state observer
Least mean squares methods
Linear systems
Mechatronics
Observers
Parameter estimation
Simulation
sliding mode observer
sliding window least square
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this paper, an observer-based online parameter estimation method for linear systems is proposed. It is shown that when the sampling interval is small enough, the sign of unmeasurable state estimation errors can be estimated and used to construct a new extended sliding mode observer with finite time convergence property. After the finite-time convergence process of the observer, a sliding window least square method (SWLS) is applied to fit estimated states while estimating unknown system parameters online. The advantages of the proposed observer-based parameter estimation method are that (i) when the sampling interval is small enough, all system states can be estimated in finite time; (ii) no recursive calculation is needed in the parameter estimation process; (iii) the simple design and implementation of the proposed estimation method is suitable for practical applications. Several simulation results are given to support the advantages and effectiveness of the proposed estimation method.
ISSN:2325-0690
DOI:10.1109/ICAMechS57222.2022.10003468