Skip to main navigation Skip to search Skip to main content

A Dynamics Modeling Method of Manipulator Based on Long Short Term Memory Neural Network

  • Yameng Zhu
  • , Hairui Zhang
  • , Guofeng Zhou
  • , Zhuo Liang
  • , Rui Lyu
  • , Yanfang Liu
  • China Acadamy of Launch Vehicle Technology
  • School of Astronautics, Harbin Institute of Technology

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

Abstract

Dynamics modeling plays an important role in studying the dynamics of manipulator. However, the traditional dynamics model is described by highly nonlinear coupled ordinary differential equations, which requires a large amount of computation and is difficult to be accurately modeled. Based on this, a dynamics modeling method of manipulator based on Long Short Term Memory (LSTM) neural network is proposed in this paper. The proposed network includes two LSTM layers and one full connection layer, which is built based on TensorFlow environment. The data sets were collected through dynamic simulations. The results show that the data curves generated by the network is basically consistent with that of the test set, and the fitting degree is high within the normal working range, which indicates that the method of building the dynamics model of the manipulator by using LSTM neural network is effective.

Original languageEnglish
Title of host publication2021 4th International Conference on Mechatronics, Robotics and Automation, ICMRA 2021
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages102-107
Number of pages6
ISBN (Electronic)9781665427272
DOIs
StatePublished - 2021
Externally publishedYes
Event4th International Conference on Mechatronics, Robotics and Automation, ICMRA 2021 - Virtual, Zhanjiang, China
Duration: 22 Oct 202124 Oct 2021

Publication series

Name2021 4th International Conference on Mechatronics, Robotics and Automation, ICMRA 2021

Conference

Conference4th International Conference on Mechatronics, Robotics and Automation, ICMRA 2021
Country/TerritoryChina
CityVirtual, Zhanjiang
Period22/10/2124/10/21

Keywords

  • Dynamics model
  • LSTM
  • Long Short Term Memory
  • Manipulator
  • Neural network

Fingerprint

Dive into the research topics of 'A Dynamics Modeling Method of Manipulator Based on Long Short Term Memory Neural Network'. Together they form a unique fingerprint.

Cite this