Sim2Real for joint robotic locomotion and manipulation with RCAN
This repo complements this project of robotics learning by providing Sim2Real transfer (from simulation to real-world) of the policy learned in simulation. The method used for Sim2Real borrows the idea of RCAN and extends it to locomotion and simultaneous manipulation and locomotion in arbitrary scenes.
Dataset generation
Validation results (in simulation)
Real-world results
