A fast test and compensation system for optical encoders based on extreme learning machine-Fourier Neural Network

Jiachen Zhao, Jie Chen, Fang Deng, Hongda Li

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

2 Citations (Scopus)

Abstract

This paper describes fast test and compensation system for optical encoders. The resolution of system is 0.375', and it can measure the actual accuracy of an optical encoder in four minutes. Furthermore, a method called Extreme Learning Machine-Fourier Neural Network (ELM-FNN) are proposed to compensate the error. Fourier neural network (FNN) is chosen to fit the curve of the optical encoder's output, and the weights of FNN is calculated by Extreme Learning Machine (ELM). Experimental results demonstrate that ELM-FNN effectively improve the accuracy of the optical encoder. Compared to a back propagation neural network (BP net) and a standard FNN, ELM-FNN has advantages of higher accuracy and less training time.

Original languageEnglish
Title of host publicationProceedings - 2017 32nd Youth Academic Annual Conference of Chinese Association of Automation, YAC 2017
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages78-82
Number of pages5
ISBN (Electronic)9781538629017
DOIs
Publication statusPublished - 30 Jun 2017
Event32nd Youth Academic Annual Conference of Chinese Association of Automation, YAC 2017 - Hefei, China
Duration: 19 May 201721 May 2017

Publication series

NameProceedings - 2017 32nd Youth Academic Annual Conference of Chinese Association of Automation, YAC 2017

Conference

Conference32nd Youth Academic Annual Conference of Chinese Association of Automation, YAC 2017
Country/TerritoryChina
CityHefei
Period19/05/1721/05/17

Keywords

  • Compensation
  • Extreme Learning Machine
  • Fourier Neural Network
  • Optical encoder

Fingerprint

Dive into the research topics of 'A fast test and compensation system for optical encoders based on extreme learning machine-Fourier Neural Network'. Together they form a unique fingerprint.

Cite this