After I finished all my exams and took a rest for a couple of weeks I returned to my master thesis.
So what did I do in this period:
- added separate neural networks that acts as forward and inverse model, which both learns from data gathered during CACLA learning (to reach object).
- created a C++ class with a lot of helpful commands for iCub (methods for environment construction, body manipulation, etc.)
- created class which specializes on grasping (hand manipulation) with parametrized methods to hand flexion, extension and such things.
I plan to connect this grasping class to some learning algorithm (like CACLA) and experiment with it.
I also read more carefully an article about Infant grasp learning and found out lots of interesting biological information (maybe I can later compare my experiment with iCub to how human infants learn to grasp).
December 2017 M T W T F S S « Mar 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
iCub on Twitter