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. Instead, the output of a Lagrangian Neural Network (LNN) is limited by physical constraints and hence is expected to predict the (inverse) multibody system behaviour more robustly. We propose to generate the feedforward control by first training a LNN that captures already most features in experimental data and next add a FFN to account for a relatively small residual. Experimental results from a fully actuated 2-DOF manipulator with flexure joints show that the accuracy of the controlled motion using this approach is comparable to using an identified inverse plant model built from the system’s equations of motion.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Abdul-hadi, O.: Machine learning applications to robot control. PhD thesis, University of California, Berkeley (2018)
Agrawal, A., et al.: TensorFlow eager: a multi-stage, python-embedded DSL for machine learning. In: Talwalkar, A., Smith, V., Zaharia, M. (eds.) Proceedings of Machine Learning and Systems, vol. 1, 178–189 (2019)
Brouwer, D.M., Folkersma, K.G.P., Boer, S.E., Aarts, R.G.K.M.: Exact constraint design of a two-degree of freedom flexure-based mechanism. J. Mech. Rob. 5(4), 041011 (2013)
Filip, S., Javeed, A., Trefethen, L.N.: Smooth random functions, random ODEs, and Gaussian processes. SIAM Rev. 61(1), 185–205 (2019)
Heerze, E.: Data-driven feedforward control of a multi degree of freedom manipulator with flexure joints using machine learning. MSc thesis, University of Twente, Enschede (2021)
Khalil, W., Dombre, E.: Modeling. Identification and Control of Robots. Kogan Page Science, London (2002)
Liu, Z., Wang, B., Meng, Q., Chen, W., Tegmark, M., Liu, T.-Y.: Machine-learning nonconservative dynamics for new-physics detection. Phys. Rev. E 104, 055302 (2021)
Lutter, M., Ritter, C., Peters, J.: Deep Lagrangian Networks: using physics as model prior for deep learning. In: 7th International Conference on Learning Representations (ICLR) (2019)
Lutter, M., Peters, J.: Combining physics and deep learning to learn continuous-time dynamics models. arXiv:2110.01894 (2023)
Wu, J., Wang, J., You, Z.: An overview of dynamic parameter identification of robots. Rob. Comput.-Integr. Manuf. 26(5), 414–419 (2010)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2024 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Heerze, E., Rosic, B., Aarts, R. (2024). Feedforward Control for a Manipulator with Flexure Joints Using a Lagrangian Neural Network. In: Nachbagauer, K., Held, A. (eds) Optimal Design and Control of Multibody Systems. IUTAM 2022. IUTAM Bookseries, vol 42. Springer, Cham. https://doi.org/10.1007/978-3-031-50000-8_12
Download citation
DOI: https://doi.org/10.1007/978-3-031-50000-8_12
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-49999-9
Online ISBN: 978-3-031-50000-8
eBook Packages: EngineeringEngineering (R0)