CHI 2026 best paper uses EMS and embodied AI to guide physical tasks
MacrinePhD · x · 2026-07-22
Researchers built an embodied AI system that tracks body pose, vision, and location to guide people through physical tasks in real time using electrical muscle stimulation (EMS).
- Instead of a fixed script, the system infers the next action and coordinates muscles step by step.
- The demo includes actions such as flexing the biceps to reach, stimulating the fingers to grip, and activating the triceps to lower a bike rack.
- The poster says the work won Best Paper at CHI 2026 and presents it as an example of AI moving into physical assistance, not just text or video.
- Code and details are linked, along with a demo video.
Related event: CHI 2026 Best Paper: Embodied AI Guides Human Actions via EMS(3 posts)→
More from Embodied
- China’s driverless delivery trucks are already reshaping logistics — CurieuxExplorer · 2026-07-22
- See2Act teaches robots where to look and how to act in one denoising loop — heghbalz · 2026-07-22
- Humanoid robot fight shows may already out-earn many nine-figure startups — kscottz · 2026-07-22
- AlayaRenderer-Flash lifts a generative world renderer from 0.56 FPS to 31.54 FPS — AlayaLab · 2026-07-22
- Autonomous trucks are cutting logistics costs across China’s highways, ports and mines — TansuYegen · 2026-07-22
- A Real Steel-style robot moment is now being framed as reality in China — CurieuxExplorer · 2026-07-22