Adaptive Estimation of Time-Varying Parameters using DREM

Emil Lykke Diget; Christoffer Sloth

doi:10.1109/CDC49753.2023.10383699

Adaptive Estimation of Time-Varying Parameters using DREM

Emil Lykke Diget^*, Christoffer Sloth

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

87 Downloads (Pure)

Abstract

In this paper we present a method for estimating time-varying parameters in a linear regression equation. We combine local polynomial regression with dynamic regressor extension and mixing to independently estimate the parameters. During local polynomial regression, a time-varying parameter is approximated by locally constant polynomial coefficients. We propose to use the Bernstein basis instead of the commonly used monomial basis to improve numerical conditioning. A simulation example shows that our proposed estimator has improved performance compared to a similar method and allows a higher polynomial order.

Original language	English
Title of host publication	2023 62nd IEEE Conference on Decision and Control (CDC)
Publisher	IEEE
Publication date	Dec 2023
Pages	3186-3191
ISBN (Electronic)	979-8-3503-0124-3
DOIs	https://doi.org/10.1109/CDC49753.2023.10383699
Publication status	Published - Dec 2023
Event	2023 62nd IEEE Conference on Decision and Control (CDC) - Singapore, Singapore Duration: 13. Dec 2023 → 15. Dec 2023

Conference

Conference	2023 62nd IEEE Conference on Decision and Control (CDC)
Country/Territory	Singapore
City	Singapore
Period	13/12/2023 → 15/12/2023

Series	Proceedings - IEEE Conference on Decision and Control
ISSN	0743-1546

Documents & Links

10.1109/CDC49753.2023.10383699

Open Access VersionAccepted manuscript, 203 KB

SDU Link

Check access to the material in SDU Library's search engine

1 Ph.D. thesis

Kinesthetic Teaching for Robotic Assembly
Diget, E. L., 14. Oct 2024, Syddansk Universitet. Det Tekniske Fakultet. 180 p.
Research output: Thesis › Ph.D. thesis

Open Access
File
127 Downloads (Pure)

1 Finished
1 Active

EUROfusion BB RH DEMO (Control)
Sloth, C. (Project participant), Diget, E. L. (Project participant), Santos, B. M. (Project participant), Lindvig, A. P. (Project participant), Nowak, C. (Project participant), de Nóbrega, R. (Project participant), Singh, R. (Project participant), Shen, Z. (Project participant), Petersen, H. G. (Project participant), Park, Y. (Project participant) & Larsen, M. E. (Project participant)
01/05/2024 → 31/12/2026
Project: EU
PIRAT - Programming Ignition for Robotic Assembly Tasks
Sloth, C. (Project participant)
01/10/2019 → 01/10/2023
Project: Research

Cite this

@inproceedings{7cf38d73232e4dabb14d0ac0cd72593d,

title = "Adaptive Estimation of Time-Varying Parameters using DREM",

abstract = "In this paper we present a method for estimating time-varying parameters in a linear regression equation. We combine local polynomial regression with dynamic regressor extension and mixing to independently estimate the parameters. During local polynomial regression, a time-varying parameter is approximated by locally constant polynomial coefficients. We propose to use the Bernstein basis instead of the commonly used monomial basis to improve numerical conditioning. A simulation example shows that our proposed estimator has improved performance compared to a similar method and allows a higher polynomial order.",

author = "Diget, {Emil Lykke} and Christoffer Sloth",

year = "2023",

month = dec,

doi = "10.1109/CDC49753.2023.10383699",

language = "English",

series = "Proceedings - IEEE Conference on Decision and Control",

publisher = "IEEE",

pages = "3186--3191",

booktitle = "2023 62nd IEEE Conference on Decision and Control (CDC)",

address = "United States",

note = "2023 62nd IEEE Conference on Decision and Control (CDC) ; Conference date: 13-12-2023 Through 15-12-2023",

}

TY - GEN

T1 - Adaptive Estimation of Time-Varying Parameters using DREM

AU - Diget, Emil Lykke

AU - Sloth, Christoffer

PY - 2023/12

Y1 - 2023/12

N2 - In this paper we present a method for estimating time-varying parameters in a linear regression equation. We combine local polynomial regression with dynamic regressor extension and mixing to independently estimate the parameters. During local polynomial regression, a time-varying parameter is approximated by locally constant polynomial coefficients. We propose to use the Bernstein basis instead of the commonly used monomial basis to improve numerical conditioning. A simulation example shows that our proposed estimator has improved performance compared to a similar method and allows a higher polynomial order.

AB - In this paper we present a method for estimating time-varying parameters in a linear regression equation. We combine local polynomial regression with dynamic regressor extension and mixing to independently estimate the parameters. During local polynomial regression, a time-varying parameter is approximated by locally constant polynomial coefficients. We propose to use the Bernstein basis instead of the commonly used monomial basis to improve numerical conditioning. A simulation example shows that our proposed estimator has improved performance compared to a similar method and allows a higher polynomial order.

U2 - 10.1109/CDC49753.2023.10383699

DO - 10.1109/CDC49753.2023.10383699

M3 - Article in proceedings

T3 - Proceedings - IEEE Conference on Decision and Control

SP - 3186

EP - 3191

BT - 2023 62nd IEEE Conference on Decision and Control (CDC)

PB - IEEE

T2 - 2023 62nd IEEE Conference on Decision and Control (CDC)

Y2 - 13 December 2023 through 15 December 2023

ER -

Adaptive Estimation of Time-Varying Parameters using DREM

Abstract

Conference

Documents & Links

SDU Link

Fingerprint

Related research output

Kinesthetic Teaching for Robotic Assembly

Related projects

EUROfusion BB RH DEMO (Control)

PIRAT - Programming Ignition for Robotic Assembly Tasks

Cite this