This code is a reimplementation of the guided policy search algorithm and LQG-based trajectory optimization.
This implementation is loosely based on a code provided by rll.berkeley.edu/gps and was enhanced here for industrial usage. This includes algorithm and policy improvements, support for OPC-UA communication and multiple vizualisation tools.
This project is funded by VDMA germany. For further information please visit i40-inpuls.de
See INSTALL.md for installation instructions.
python .\main.py gym_fetchreach_lqr
What to see what the robot is doing?
Set 'render'
to True
in the agent block in experiments\gym_fetchreach_lqr\hyperparams.py
.