Copied


Pegatron Implements AI-Enabled Digital Twins to Optimize Factory Operations

Alvin Lang   Jun 03, 2024 04:58 3 Min Read


In an effort to address the increasing pressures of shortening production cycles and enhancing productivity while reducing costs, Pegatron, a leading electronics manufacturer based in Taiwan, has adopted AI-enabled digital twin technology. The initiative is part of a broader trend among manufacturers to invest in industrial digitalization and advanced technologies.

Pegatron's Digital Twin Initiative

Pegatron developers are utilizing NVIDIA AI, NVIDIA Omniverse, and Universal Scene Description (USD) to create PEGAVERSE, a sophisticated digital twin platform. This platform allows engineers and factory managers to collaboratively plan, simulate, and optimize their production lines, providing real-time insights into facilities, equipment, and maintenance tasks. According to NVIDIA Technical Blog, the platform has enabled Pegatron teams to quickly identify and resolve operational issues.

Andrew Hsiao, Associate Vice President for the AI Development Division at Pegatron, highlighted the benefits of PEGAVERSE, stating, “With PEGAVERSE, we easily integrate all our resources into the virtual world, from design, optimization, maintenance, and operations. It fulfills our need for real-time visibility, predictive maintenance, process optimization, remote operations, quality assurance, risk management, innovation, and cost reduction in our manufacturing operations.”

Using Virtual Factories to Optimize Operations

The PEGAVERSE platform supports a variety of use cases, including predictive maintenance, process optimization, resource planning, remote monitoring, and quality control. Initially developed by a team of 15, the project has expanded to over 100 developers, including experts in 3D modeling, computational fluid dynamics (CFD), finite element analysis (FEA), AI, data engineering, and full-stack development.

The platform features three core components:

  1. PEGAAi: An MLOps deep learning platform with no-code and low-code capabilities that streamline and automate data collection, labeling, training, deployment, and monitoring.
  2. Large Language Models (LLMs): These models enable engineers to develop customized automation processes, adding flexibility and intelligence to their design and manufacturing workflows.
  3. Digital Twin Technologies: These technologies allow for the creation of accurate, large-scale virtual production environments used for model training and simulation.

With PEGAVERSE, Pegatron has established five virtual factories across Taiwan, India, Indonesia, China, and Vietnam. These virtual environments are used to monitor operations remotely and run simulations to identify bottlenecks and optimize production lines.

Unified Tools and Data with OpenUSD

One of the significant challenges Pegatron faced was connecting different tools and data to create real-time, physics-based models of their facilities. Using OpenUSD, an extensible framework for 3D worlds, Pegatron developers integrated various data, tools, and processes within the PEGAVERSE platform. This integration has accelerated the design, simulation, and development of automated equipment and production lines.

Hsiao emphasized the importance of OpenUSD, likening it to HTML's role in the growth of the Internet. “In the age of AI, there is a need for standards that can describe more real-world information, and OpenUSD fills this gap,” he said. “We see OpenUSD as a new form of the digital asset—it will be the foundation for establishing new processes and management principles to increase efficiency and create higher value for our customers.”

Enhancing Digital Twin Experiences with IoT and Generative AI

Pegatron has also integrated generative AI and IoT technologies into PEGAVERSE. Their in-house IoT service, PEGAAIoT, provides real-time insights, remote control, and log replay features. AI agents powered by NVIDIA NeMO enable engineers to interact with the virtual factories using natural language, facilitating quick access to critical information.

Additionally, NVIDIA Isaac Sim and NVIDIA Metropolis platforms are used to train perception AI models, simulate and fine-tune robotic systems, and develop automated optical inspection workflows. These tools help ensure the quality and efficiency of Pegatron's production lines.

By leveraging AI, Omniverse, and OpenUSD, Pegatron aims to continuously improve and transform its manufacturing operations, benefiting from real-time insights and predictive analytics provided by AI-enabled digital twins.


Read More