Nvidia's big bet on physical AI aims for safer robotaxis, humanoid robots
Nvidia is extending its Halos safety stack from autonomous vehicles to robotaxis and humanoid robots.
Nvidia is extending its Halos safety software from autonomous vehicles to robotics, including robotaxis and humanoid robots. Halos continuously monitors hardware blocks and software libraries, isolates safety-critical workloads, and uses the Holoscan Sensor Bridge to connect sensor data to processing that can identify corrupted data. The package also includes virtual-environment simulations and an inspection lab for partner feedback on safety issues. Nvidia says functional safety is relatively uniform for cars but must be customizable for robots operating in unstructured factories, warehouses, and hospitals.
- Halos monitors hardware and software and isolates safety-critical computing workloads.
- Holoscan Sensor Bridge connects sensors to processing that can flag corrupted data.
- Car safety definitions are consistent, but robot safety needs vary widely.
- Developers can add custom safety functions without breaking the underlying stack.
- Vendors
- Nvidia
- Products
- HalosHoloscan Sensor Bridge
- Organizations
- Nvidia
From the original article
1 min read at arstechnica.com
Full-stack safety solution for physical AI is being used by robotics companies.
By Jeremy Hsu · arstechnica.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.