Self-driving project spent 87.5% of time in simulation; imitation learning beat lane estimation
abursuc · x · 2026-09-18
Author shares takeaways from a self-driving project: 87.5% of the time was spent in simulation, with only a small fraction in real-world testing. On methods, imitation learning proved much more efficient than center lane estimation. Field tests in the mountains in November encountered multiple adverse weather events, highlighting how environmental conditions affect real-world data collection.
More from Embodied
- Hugging Face co-founder Thom Wolf shows off iterating robot head prototype — Thom_Wolf · 2026-09-18
- Open-source desk robot Sudo launches: self-hosted AI, $0 subscription, 25 founder units — wateriscoding · 2026-09-18
- 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