Feedforward control for a manipulator with flexure joints using a Lagrangian Neural Network

Eline Heerze, R.G.K.M. Aarts*, Bojana Rosic

*Corresponding author for this work

Research output: Contribution to conferenceAbstractAcademic

135 Downloads (Pure)

Abstract

Feedforward control of a manipulator can be generated with a sufficiently accurate stable inverse model of the manipulator. A Feedforward Neural Network (FNN) can be trained with experimental data to generate feedforward control without knowledge about the system at hand. However, the FNN output can show unphysical behaviour especially in operational regimes where the training data is sparse. We consider including a Lagrangian Neural Network (LNN) that is expected to predict the (inverse) multibody system behaviour more robustly.
Original languageEnglish
Pages1
Number of pages2
Publication statusPublished - 18 Jul 2022
EventIUTAM Symposium on Optimal Design and Control of Multibody Systems 2022: Adjoint Methods, Alternatives, and Beyond - Hamburg University of Technology (TUHH), Hamburg, Germany
Duration: 18 Jul 202221 Jul 2022
https://www.tuhh.de/mum/iutam-symposium-2022.html

Conference

ConferenceIUTAM Symposium on Optimal Design and Control of Multibody Systems 2022
Country/TerritoryGermany
CityHamburg
Period18/07/2221/07/22
Internet address

Fingerprint

Dive into the research topics of 'Feedforward control for a manipulator with flexure joints using a Lagrangian Neural Network'. Together they form a unique fingerprint.

Cite this