# Pittsburgh Team Runs Meta's Segmentation Models On-Device for a Wheelchair-Mounted Robotic Arm

The $41.5 million project, funded by the Advanced Research Projects Agency for Health (ARPA-H), keeps perception local on battery-powered hardware, removing the dependence on network connectivity that makes cloud-based inference unsafe for assistive robotics.

- Published: 2026-09-27T06:30:19.445Z
- Canonical: https://polylog.news/ai/2026-09-27/pittsburgh-team-runs-meta-s-segmentation-models-on-device-fo
- Publisher: Polylog (AI desk)
- Section: tech
- Sources: [Meta AI](https://ai.meta.com/blog/assistive-robotics-university-of-pittsburgh-sam-dino/), [Google Research](https://research.google/research-areas/google-earth-ai/)

Researchers at the University of Pittsburgh's Human Engineering Research Laboratories are running Meta's DINOv3 and Segment Anything models directly on the computing hardware carried by a robotic wheelchair and arm, according to Meta's acco…

This story is for subscribers. Read it in full at https://polylog.news/ai/2026-09-27/pittsburgh-team-runs-meta-s-segmentation-models-on-device-fo (subscription information: https://polylog.news/pricing).