Stable Model-based Control with Gaussian Process Regression for Robot Manipulators

Thomas Beckers, Jonas Umlauft, Sandra Hirche

Research output: Contribution to journalArticlepeer-review

21 Scopus citations

Abstract

Computed-torque control requires a very precise dynamical model of the robot for compensating the manipulator dynamics. This allows reduction of the controller's feedback gains resulting in disturbance attenuation and other advantages. Finding precise models for manipulators is often difficult with parametric approaches, e.g. in the presence of complex friction or flexible links. Therefore, we propose a novel computed-torque control law which consists of a PD feedback and a dynamic feed forward compensation part with Gaussian Processes. For this purpose, the nonparametric Gaussian Process regression infers the difference between an estimated and the true dynamics. In contrast to other approaches, we can guarantee that the tracking error is stochastically bounded. Furthermore, if the number of training points tends to infinity, the tracking error is asymptotically stable in the large. In simulation and with an experiment, we demonstrate the applicability of the proposed control law and that it outperforms classical computed-torque approaches in terms of tracking precision.

Original languageEnglish
Pages (from-to)3877-3884
Number of pages8
JournalIFAC Proceedings Volumes (IFAC-PapersOnline)
Volume50
Issue number1
DOIs
StatePublished - Jul 2017

Keywords

  • Adaptive system
  • Data-based control Nonparametric methods
  • Stability of nonlinear systems
  • Stochastic control
  • control
  • robotic manipulators

Fingerprint

Dive into the research topics of 'Stable Model-based Control with Gaussian Process Regression for Robot Manipulators'. Together they form a unique fingerprint.

Cite this