Interface for robot learning experiments in ROS. Based on gym environments
This package (along with https://github.com/juancamilog/kusanagi) was used to generate the results of our IROS 2018 submission " Synthesizing Neural Network Controllers with Probabilistic Model based Reinforcement Learning"
A video demonstration of some of the results in the paper is available here
To run the aqua robot simulator experiments, we recommend to use the docker compose file available in https://github.com/juancamilog/aqua_robot_learning_docker
This is work in progress, and I'll be updating the install instructions in the coming months. If you have any questions on how to run it, please write me an e-mail at gamboa [at] cim [dot] mcgill [dot] ca