The summary of my work on the thesis during the last 2 months.
- I read a lot of articles and books about Reinforcement Learning and its (C)ACLA specialized algorithms. I studied pseudocodes and also some C++ implementations of these algorithms (particularly for the maze [discrete space] problem and the cart pole [continuous space] problem).
- To get the best background in RL as possible I watched several lectures of Machine Learning at Stanford University.
- I made some C++ programmes for the iCub after I had studied the tutorial section which is available in the iCub software repository. I ran all the tutorial applications and precisely studied their source codes.
- I was particularly focused on the Cartesian Interface Tutorial which shows how to control a robot’s limb in the operational space. I would like to use the Cartesian module in the first phase of my thesis because it provides coordinates of the end-effector (hand) so I can easily implement the reward function of the RL algorithm.
January 2018 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