This is a project for RL training infrastructure for uwrt arm movement as part of the UWRobotics club, which I collaborated with Melvin Wang.
The following diagram provides an overview of our project. I also worked with Melvin to develop our pipeline with traditional controls for keyboard typing, see our main UWRobotics repo here.
I've also included slides for a recent presentation I gave on an overview of our project in samples/IntroToAppliedRL.pdf.
We implmented the following packages/repos for this work OpenAI Gym (Training Env) + PyBullet (Simulator) + Stable-baselines3 (RL Algorithmm).
Check out our 2021 SAR submission for this year's (cancelled) URC competition here.
Here is a sample of our robot typing on a keyboard in simulation!
