Each robot runs two complementary pretrained components.
Semantic messaging helps distributed robots coordinate long manipulation tasks
DuoMind pairs high-level vision-language reasoning with low-level action policies on each robot, allowing peers to coordinate through natural-language messages without a central controller.
Big Tech
Hanchu Zhou · Dechen Gao · Hang Wang · Brendan Lynch · Boqi Zhao · Qiyao Ma · +2 more
University of California, Davis · Microsoft Research · Analog Devices
Research Digest··2 min read
Zhou and colleagues present DuoMind, a hierarchical system in which each robot independently plans, communicates with peers and executes detailed actions.
Why this paper
From Microsoft Research and 2 others
In one line
DuoMind uses hierarchical VLM orchestrator and VLA action model with semantic communication to improve multi-robot coordination performance.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ·No stated limitations found
- ·No benchmark numbers found
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§