CRoKoDiLe 🐊

Cross-Robot Knowledge Transfer in Deep Reinforcement Learning

This project addresses the challenge of transferring knowledge between two different industrial robots, i.e. different kinematics.

The proposed approach is executed in four stages:

1. Generate expert knowledge
2. Generate domain mappings
3. Transfer expert knowledge to target domain
4. Boost target agent