This repo contains a suite of OpenAI gym-wrapped RL environments, that have similar problem properties than the Parkinson's disease specific aDBS control problem. These environments are developed based on the FlappyBird implementation provided in the PyGame Learning Environment (PLE) library.
See the repo FlappyBird_agents/examples for details on how to use the environments.
FlappyBird is adapted to have non-stationary dynamics and a noisy reward.
If you are having issues, please let me know. Contact me at: [email protected]