Universal Robots Safety-Critical Edge AI Platform Integration for Industrial Applications

[Erkan Teskancan]

## Integration of Safety-Critical Edge AI Platforms in Industrial Applications

QNX, a division of BlackBerry, announced an expanded collaboration with NVIDIA at Hannover Messe. The integration of QNX® OS for Safety 8.0 with the NVIDIA IGX Thor platform enables the development of safety-critical edge AI systems in regulated environments such as surgical robotics, autonomous mobile robots (AMR), and industrial automation.

### Architectural and Technical Approach
This integration addresses the need for safety-critical control systems and high-performance artificial intelligence operations to run on a single hardware platform. QNX OS for Safety 8.0 combined with the NVIDIA IGX Thor Halos Safety Stack offers a unified system architecture that supports mixed-criticality tasks. AI-based perception, spatial planning, and decision-making functions run on NVIDIA's accelerated cores, while critical safety functions operate in an isolated and deterministic manner within the QNX microkernel.

### Technical Reliability and Certification
Functional safety and system reliability are mandatory in sectors such as medical technology and heavy industry. The unified platform supports regulatory-compliant certification processes. The QNX microkernel architecture prevents low-priority processes from interfering with high-priority safety tasks, thereby maintaining system integrity. This structure is critical for software-defined machines where autonomous behaviors are managed at the edge.

### Technology Transfer from Automotive to Industry
QNX and NVIDIA are standardizing the safety architecture, previously proven with the NVIDIA DRIVE AGX Thor development kit, by transferring it to industrial and medical applications. Thus, the highly deterministic and functional safety standards applied in autonomous vehicles are now also being applied to smart factory equipment and digital supply chain robots.

### Development and Early Access
The unified platform is designed to support the entire product lifecycle, from early prototype stages to mass production. This compatible software architecture streamlines the development process, accelerating the time to market for complex AI systems. Early access registration for the NVIDIA IGX Thor Developer Kit and QNX has been opened, offering selected customers the opportunity to develop and validate safety-critical applications.

### Features of NVIDIA IGX Thor and QNX OS for Safety 8.0

• Deterministic microkernel real-time operating system (RTOS)
• High-performance AI operations with NVIDIA accelerated computing cores
• Halos Safety Stack supporting functional safety features and certification
• Coexistence of safety-critical and AI tasks in an isolated architecture
• Compliance with industrial and medical regulations
• Early access and prototyping support with development kits

This integration is considered a significant step forward for the future of safety and artificial intelligence in industrial automation and robotics.