TR2018-201

Derivative-free online learning of inverse dynamics models


    •  Romeres, D., Zorzi, M., Camoriano, R., Traversaro, S., Chiuso, A., "Derivative-free online learning of inverse dynamics models", IEEE Transactions on Control Systems Technology, DOI: 10.1109/​TCST.2019.2891222, March 2019.
      BibTeX TR2018-201 PDF
      • @article{Romeres2019mar,
      • author = {Romeres, Diego and Zorzi, Mattia and Camoriano, Raffaello and Traversaro, Silvio and Chiuso, Alessandro},
      • title = {Derivative-free online learning of inverse dynamics models},
      • journal = {IEEE Transactions on Control Systems Technology},
      • year = 2019,
      • month = mar,
      • doi = {10.1109/TCST.2019.2891222},
      • url = {https://www.merl.com/publications/TR2018-201}
      • }
  • MERL Contact:
  • Research Areas:

    Data Analytics, Optimization

Abstract:

This paper discusses online algorithms for inverse dynamics modelling in robotics. Several model classes including rigid body dynamics (RBD) models, data-driven models and semiparametric models (which are a combination of the previous two classes) are placed in a common framework. While model classes used in the literature typically exploit joint velocities and accelerations, which need to be approximated resorting to numerical differentiation schemes, in this paper a new “derivative-free” framework is proposed that does not require this preprocessing step. An extensive experimental study with real data from the right arm of the iCub robot is presented, comparing different model classes and estimation procedures, showing that the proposed “derivative-free” methods outperform existing methodologies.