RSTAR overcoming obstacles using DRL based on energy reward function

We taught our reconfigurable RSTAR robot to overcome obstacles and the results are amazing. The DRL is based on an energy reward function allowing it to shorten learning time and improve results.
By: Or Simhon, Zohar Karni, Sigal Berman, and David Zarrouk
ieeexplore.ieee.org/document/...

Thank you for sharing.