UBC student deploys RL agent on robotic air hockey table

View on Reddit

The two-year undergrad capstone from Engineering Physics at UBC trained a reinforcement learning policy in simulation and deployed it directly onto a custom-built robotic air hockey table. The system includes computer vision and direct sim-to-real transfer.