UC Berkeley's ATLAS drives humanoid robots from a single human video, no VLA or task-specific training
ChongZzZhang · x · 2026-11-13
ATLAS, a UC Berkeley project with NVIDIA, HKU and Stanford, won Best Demo at the CoRL 2026 Agentic Robotics Workshop. From one human video demonstration and a single prompt—no robot demos, no task-specific sim training, no teleoperation, no VLA—it gets a real humanoid to do whole-body tasks like descending stairs, picking up a chip bag, and collecting trash.
The architecture splits roles: System 2 is a VLM agent that knows the task, System 0 is a whole-body motion tracker, and an agent-written harness links the two, learning at test time how this specific robot executes the command in this scene. All sensing is onboard.
More from Embodied
- NVIDIA's VLA driving model hits 20 decisions/s on Jetson Thor after Embedl optimization — No-Dragonfly6246 · 2026-11-13
- Roboticist: You Can't Do Robotics Without NVIDIA's Stack; Jensen's Long Bets Are Underrated — tianyin_xu · 2026-11-13
- US humanoids span pilots, early commercial expansions and production ramps amid FCC import limits — ingliguori · 2026-11-13
- Invictus Robotics unveils Aeras, a force-controlled aerial manipulation drone for industrial work — lukas_m_ziegler · 2026-11-13
- Zero-HLM: 45 sim demos become 4,800, humanoid zero-shot sim2real hits 69% vs 23% baseline — 机器之心 · 2026-11-13
- RoboScience launches Visics-Agent robot assistant with 78% success on long-horizon L3 tasks — 机器之心 · 2026-11-13