sudhirpratapyadav / Vision-UR5-Deep-RL

Vision-based-control-of-UR5-robotic-arm-using-Deep-RL

This project tries to solve Visual Servoing problem using Reinforcement Learning. UR5 robotic arm is trained to visually servo a speicific color object via Reinforcement Learning.

Please read documentation to know more details of project.

Contributions by Open Source Libraries

In DDPG Code 3 Folders/Branches are taken from open source projects of other people. I am very greateful towards them for that. These are following:

1. Deep Mind Atari Deep Q Learner -- Please check README.md of that branch

2. Torch Policy Gradient -- Timothy P. Lillicrap, Jonathan J. Hunt, Alexander Pritzel, Nicolas Heess, Tom Erez, Yuval Tassa, David Silver, Daan Wierstra, Implemented by Yannis M. Assael -- Please check README.md of that branch

3. 3D Model of UR5 robot -- Basic model was taken from open source library -- Some upgrades have been made to model and Gazeebo World Model to suit this specific project.