Industrial Robotics Simulation Platform for Physical Artificial Intelligence Applications

[Cengiz Özemli]

## Industrial Robotic Simulation Platform for Physical AI Applications

ABB Robotics and NVIDIA are collaborating to integrate NVIDIA Omniverse simulation libraries into the ABB RobotStudio engineering environment to support the development and deployment of AI-powered industrial robots. This partnership aims to increase the reliability of robot training in the digital environment and ensure the seamless transfer of developed models to real production processes.

### Scope of Collaboration

Industrial robots are trained based on simulated AI models before being deployed in real production environments. However, the differences between simulation and the real world, known as the "sim-to-real gap," have limited the reliability of virtual training. ABB Robotics develops industrial robots and engineering software used in production automation, while NVIDIA provides high-performance computing and simulation platforms. This collaboration combines robotic engineering tools with advanced simulation infrastructure to support scalable production automation and physical AI applications.

### Technical Solution and Tasks

• ABB provides the RobotStudio engineering platform. The virtual controller within this platform runs on the same software used in physical robots. This allows control programs tested in simulation to be used directly on production equipment in the same way.
• NVIDIA offers accelerated computing technologies with Omniverse simulation libraries that accurately mimic lighting, materials, textures, object geometry, and environmental conditions encountered in production environments with high physical fidelity.
• The joint system is called RobotStudio HyperReality and enables engineers to generate high volumes of synthetic data for AI training.
• ABB's Absolute Accuracy technology reduces the positioning error margin between simulation and real operation from 8–15 mm to approximately 0.5 mm, achieving up to 99% accuracy.

### Implementation and Rollout

The RobotStudio HyperReality platform is planned to be made available to ABB's global RobotStudio users in the second half of 2026. The platform will be compatible with existing ABB robot controllers and engineering workflows. Additionally, ABB is evaluating the integration of the NVIDIA Jetson edge computing platform into the OmniCore robot controller architecture. This will enable robots to perform real-time AI inference on the production floor.

### Application Areas and Use Cases

• The platform is targeted for use in sectors requiring high precision, such as consumer electronics manufacturing, automotive assembly, and logistics automation.
• Foxconn is piloting the platform in consumer electronics assembly. Here, robots process small parts and prepare for various product variants using simulated training data.
• US-based company WORKR is deploying robotic solutions for small and medium-sized manufacturers using this platform.

### Results and Expected Impacts

Physically accurate simulation and synthetic training data allow manufacturers to design and validate robotic systems before physical installation. According to ABB analyses, virtual commissioning processes are accelerated by up to 80%, development costs are reduced by up to 40%, and the production speed of complex products increases. The integration of robotic engineering tools with high-fidelity simulation and AI training capabilities will support the scalable deployment of intelligent robotic systems in production environments.