search

sjauhri / Interactive-Learning-in-State-space

Implementation of a novel interactive learning method: TIPS (Teaching Imitative Policies in State-space). Part of the Master Thesis: "Interactive Learning in State-Space" at the Delft University of Technology
9 stars 2 forks source link

Request for the code of the robot part of your paper "Interactive Imitation Learning in State-Space" #4

Open Marcelbgit opened 3 years ago

Marcelbgit commented 3 years ago

Dear Snehal,

thanks for sharing your great and extremely interesting research achievement. This code here is related to simulation in OpenAI Gym environments for CartPole, LunarLanderContinuous and Reacher. Far more exciting to me is the physical part with the 7-axis robot regarding limits of learned activities and possible improvements. I would like to try this model myself for other tasks. Is it possible to provide the code for this? It would be great to follow up on such a great work.

Best regards, Marcel

sjauhri commented 3 years ago

Hello Marcel,

For the robot, we used only two joints for the tasks and used an underlying ros control iiwa stack. You can find the code in the legacy branch of this repository at https://github.com/sjauhri/Interactive-Learning-in-State-space/tree/legacy/KUKA_Task/ros_ws/src/tips.

Let me know if you have questions or need assistance.