Online semi-parametric learning for inverse dynamics modeling
Authors:
- Diego RomeresDRDiego Romeres
- Mattia ZorziMZMattia Zorzi
- Raffaello CamorianoRCRaffaello Camoriano
- Alessandro ChiusoACAlessandro Chiuso
Preprint
Abstract
This paper presents a semi-parametric algorithm for online learning of a robot inverse dynamics model. It combines the strength of the parametric and non-parametric modeling. The former exploits the rigid body dynamics equa- tion, while the latter exploits a suitable kernel function. We provide an extensive comparison with other methods from the literature using real data from the iCub humanoid robot. In doing so we also compare two different techniques, namely cross validation and marginal likelihood optimization, for estimating the hyperparameters of the kernel function.Keywords
All Related Versions
This publication has 9 references indexed in Scilit:
- Tuning complexity in regularized kernel-based regression and linear system identification: The robustness of the marginal likelihood estimatorAutomatica, 2015
- iCub Whole-Body Control through Force Regulation on Rigid Non-Coplanar ContactsFrontiers in Robotics and AI, 2015
- Rational approximations of spectral densities based on the Alpha divergenceMathematics of Control, Signals, and Systems, 2013
- On-line Dynamic Model Learning for Manipulator ControlIFAC Proceedings Volumes, 2012
- Incremental online sparsification for model learning in real-time robot controlNeurocomputing, 2011
- The iCub humanoid robot: An open-systems platform for research in cognitive developmentNeural Networks, 2010
- A Survey on Transfer LearningIEEE Transactions on Knowledge and Data Engineering, 2009
- Model Learning with Local Gaussian Process RegressionAdvanced Robotics, 2009
- Numerical Methods for Least Squares Problems1996