This module allows to carry out a two-finger stable grasp by means of a controller developed using reinforcement learning. First, the fingers move towards the object and stop as soon as a contact is detected, then the hand configuration changes to improve stability.
numpy
(on Linux typeapt-get install python-numpy
)
- run
skinManager
- run
plantIdentification --whichHand <hand> --icub <icub>
, where:<hand>
is the hand to use (left/right)<icub>
is the iCub to use (iit_black/iit_purple/tum_darmstadt)
- run
python gpControllerStableGrasp.py