Meet Halos, NVIDIA's new safety system for autonomous vehicles
What's the story
NVIDIA has unveiled Halos, a complete safety system for the development and deployment of autonomous vehicles (AVs).
The new solution combines NVIDIA's hardware, software tools, and AI research into a single framework to improve safety of self-driving vehicles.
It is designed to address the increasing demand for advanced safety in autonomous systems by combining platform, algorithmic, and ecosystem safety measures.
Halos features a safety-assessed system-on-a-chip (SoC) with hundreds of built-in safety mechanisms.
Tech integration
Halos leverages NVIDIA's advanced technologies for safety
Halos integrates solutions for AI training, simulation, and deployment by harnessing NVIDIA's DGX, Omniverse, Cosmos, and DRIVE AGX technologies.
The system also features NVIDIA DriveOS - a safety-certified operating system and DRIVE AGX Hyperion - a hardware platform that connects sensors, processors, and software.
This holistic approach promises safe data handling and unbiased training datasets while enabling automated safety evaluations via a data-driven feedback loop.
Collaboration
AI systems inspection lab supports Halos
Along with Halos, NVIDIA has also launched the AI Systems Inspection Lab - a program aimed at integrating functional safety, cybersecurity, and regulatory compliance.
It has already garnered support from industry leaders like Ficosa, OMNIVISION, Onsemi, and Continental.
With more than 1,000 patents filed and over 240 research papers published, NVIDIA continues to innovate to enable a future with reliable and safer autonomous vehicles.