Meet SENTINEL: Turning Words into Whole-Body Robot Motion
What’s new
Meet SENTINEL, an end-to-end model that lets humanoid robots turn plain-language instructions into whole-body motion - no hand-crafted pipelines or teleop in the loop.
How it works
- Trains on a large dataset built by tracking human motions in simulation and pairing them with text descriptions.
- Directly maps your words and the robot's own body signals (proprioception) to low-level motor actions.
- Uses flow matching to propose short "action chunks," then refines them with a residual action head for real-world stability.
- Supports multimodal cues by converting different inputs into text.
Why it matters
- Tighter alignment between what we say and how robots move.
- Stable execution across simulation and real hardware.
- A step toward general-purpose humanoids that can learn new tasks from language.
Paper: https://arxiv.org/abs/2511.19236v1
Paper: https://arxiv.org/abs/2511.19236v1
Register: https://www.AiFeta.com
Robotics AI Humanoid EmbodiedAI LanguageToAction HRI MachineLearning
