Driving policy trained in CARLA via imitation deployed directly on real car

abursuc · x · 2026-09-18

Continuing the #ssad2026 demo: the team trained a driving policy via imitation learning in CARLA (using a privileged map-aware expert for data) and deployed it directly on the real car upon arrival — a sim2real pipeline.

Related event: CARLA-Trained Driving Policy Deployed Directly to Real Vehicles(2 posts)→

Original post →

More from Embodied

Embodied channel →