CARLA driving expert with privileged maps trains vision-based imitation policy
abursuc · x · 2026-09-18
Presented at #ssad2026: a hybrid-vision driving model takes camera, GPS, ego state and driving actions as input to predict depth, semantic segmentation and waypoints. Training data is generated via a privileged driving expert with map access in CARLA, used for imitation learning.
Related event: CARLA-Trained Driving Policy Deployed Directly to Real Vehicles(2 posts)→
More from Embodied
- LAIA dataset: 15 hours of CARLA driving with human gaze labels for explainable end-to-end AV research — abursuc · 2026-09-18
- Figure runs humanoid robots in 30 rented Bay Area homes with no new training; predicted a $10T company — LinusEkenstam · 2026-09-18
- Vintage robotics video: a handful of PhDs testing modular architectures in the wild — abursuc · 2026-09-18
- Robotics startup founder: hosting events self-selects the hard-to-find ML talent — DominiqueCAPaul · 2026-09-18
- Developers embrace open-source physics simulator MuJoCo for modular robot prototyping — Obijuan_cube · 2026-09-18
- Stanford's FAMOS infers 3D articulation from sparse partial point clouds feed-forward — Stanford-University · 2026-09-18