Objectives

Using both, tactile and visual feedback, a robot should learn to operate various kinds of switches, knobs, etc. This comprises i) to position the hand relative to the target object utilizing characteristic tactile features on the object and its environment, and ii) to apply a characteristic motion and force profile to operate the switch. This objective will be used also as testbed to compare spiking vs. classical methods for slip detection and force profile analysis.

Expected Results

Behavioral skills allowing a robot hand to reliably operate various switches

Planned Secondments

  • IIT-iCub:
    to learn about and implement spiking methods for tactile feature detection
  • USFD:
    to learn predictive models
  • PAL:
    to evaluate methods on a different robot hand (generalization ability)

Enrolments (in Doctoral degree/s)

Bielefeld University

Supervisors

R. Haschke, C. Bartolozzi, T. Prescott, S. Terreri

Tags

ART ROB
ESR10: Learning to find and operate switches
This project is funded by the EU Horizon 2020 research and innovation programme under grant agreement No 813713 (NeuTouch)
© 2018 Neutouch
Call closed

Please note that this call has been closed therefore it will not be possibile to apply for this Research Theme.