A Regularization Ensemble Based on Levenberg–Marquardt Algorithm for Robot Calibration

Xin Luo*, Zhibin Li, Long Jin, Shuai Li

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingChapter or book articleScientificpeer-review

Abstract

This chapter investigates six regularization schemes, such as L1, L2, dropout, elastic, log, and swish. Then, an efficient ensemble incorporates six regularizations to achieve high calibration accuracy. Firstly, Sect. 5.1 discuss the research background of robot calibration. In Sect. 5.2, we introduce six regularized robot calibration schemes and the principle of an ensemble. Then, Sect. 5.3 presents experiments for the proposed ensemble. Lastly, conclusions and future work are summarized in Sect. 5.4.
Original languageEnglish
Title of host publicationRobot Control and Calibration
Subtitle of host publicationInnovative Control Schemes and Calibration Algorithms
Place of PublicationSingapore
PublisherSpringer
Pages69-89
ISBN (Electronic)978-981-99-5766-8
ISBN (Print)978-981-99-5765-1
DOIs
Publication statusPublished - 2023
MoE publication typeA3 Part of a book or another research book

Publication series

SeriesSpringerBriefs in Computer Science
VolumePart F1465
ISSN2191-5768

Keywords

  • Ensemble
  • Kinematic parameters
  • Levenberg-Marquardt algorithm
  • Overfitting
  • Regularization scheme

Fingerprint

Dive into the research topics of 'A Regularization Ensemble Based on Levenberg–Marquardt Algorithm for Robot Calibration'. Together they form a unique fingerprint.

Cite this