Hierarchical Control Architecture for Embodied Robotics Revealed
CyberRobooo · x · 2026-07-17
This content outlines a hierarchical architecture for embodied robotic systems, emphasizing the synergy of Sys0-Sys1-Sys2 to achieve hand-eye-brain coordination:
- Sys0: Low-level control operating at the 1kHz level, handling precise and stable actions like folding clothes or arranging tableware.
- Sys1: Edge computing paired with the Thor chip, managing real-time perception and reasoning in dynamic environments like supermarkets.
- Sys2: Cloud models responsible for higher-level planning, such as assisting with checkout payments.
The post also notes that Dexmal in Beijing has introduced the wheeled humanoid robot Apex, currently training in real-world scenarios:
- Folding and hanging clothes
- Assisting with store payments
- Supermarket restocking and sorting
- Arranging tableware in restaurants
The author believes this modular design helps models better connect with physical bodies, and the resulting real-world operational data will feed back into model improvement.
More from Embodied
- NVIDIA open-sources FLORA, a robotics data engine ~125x cheaper than GPT-6 Astra — nvidia · 2026-10-27
- Embodied AI startups need a different operational discipline, investor says — ___Mufasaa · 2026-10-27
- π0.5 study: raising demos from 10 to 438 lifts Place Ball in Bowl success from 7% to 83% — stingning · 2026-10-27
- GPT-6 Astra controls dexterous hands zero-shot at just 5.2% success in 500 trials — stingning · 2026-10-27
- HandsOn benchmark: 66 verified tasks and 18,000 human demos for dexterous manipulation — stingning · 2026-10-27
- DexPort: a handheld interface with a real robot hand for dexterous-hand data collection — pabbeel · 2026-10-27