The basic workspace for reinforcement learning with CoppeliaSim (VREP) simulation environments, including some demonstrated project for beginners. Some tutorials can be found at:
- Zhihu: https://zhuanlan.zhihu.com/p/398874515
- WeChat: https://mp.weixin.qq.com/s/id7fw0eGBtLqEqr-zaKKPQ
All the demos have been tested on Ubuntu 20.04, with Anaconda python environment.
You have to install the following softwares and environments for this project, the recommend operating system is Ubuntu.
- CoppeliaSim 4.2 (https://www.coppeliarobotics.com/)
- Python 3.6+ (Anaconda is recommended)
- Gym (https://github.com/openai/gym)
- Stable-baselines3 (https://github.com/DLR-RM/stable-baselines3)
- Pytorch (https://pytorch.org/)
- Visdom (pip install visdom)
- Step 1: run CoppeliaSim, import cart_pole.ttt
- Step 2: run visdom in your terminal, open your browser, and visit link: localhost:8097
- Step 3: run the script named 'demo_cart_pole_learning.py' in the sub-path ./examples
Then we have: