Modular edge AI platform for industrial automation

Industrial automation systems are increasingly required to process large data volumes locally, driven by applications such as machine vision, predictive maintenance, and autonomous systems. Conventional PLC and IPC architectures are often insufficient for these tasks, particularly when low latency and data sovereignty are required. The presented platform addresses these requirements by combining classic industrial expandability with high-performance CPU and GPU computing, positioning it as a building block for edge AI and the digital supply chain.

Compact design with full industrial expandability
Traditional 4U industrial PCs have been widely used where multiple expansion cards and high computing power are required. The new system replaces these bulky enclosures with a significantly smaller footprint while retaining up to seven PCIe/PCI expansion slots. This allows continued use of existing I/O, communication, motion control, or image acquisition cards, reducing migration effort while lowering installation space and system power consumption.

Processor and memory architecture
The platform supports Intel processors from the 12th, 13th, and 14th generations based on the Alder Lake architecture, with configurations up to Core i9 class CPUs. Combined with the Intel Q670 chipset, this enables high-bandwidth I/O and fast memory access via NVMe and SSD storage. These features support data-intensive workloads typical of industrial analytics and AI inference, delivering performance levels comparable to entry-level server systems.

GPU-enabled edge AI without active cooling
A defining technical characteristic is support for dedicated GPUs with power consumption up to 115 W in a fully fanless chassis. Heat is dissipated through solid cooling fins and direct thermal conduction, eliminating moving parts. This design increases system reliability and allows silent operation in harsh environments. GPU acceleration enables local execution of AI-based image processing, deep learning inference, and predictive maintenance algorithms, reducing latency and dependence on cloud connectivity within an automotive data ecosystem or factory network.

Industrial robustness and connectivity
The system complies with EN 50155 for shock and vibration resistance and EN 45545-2 fire protection requirements, qualifying it for safety-critical and mobile applications. Networking is provided via nine Gigabit Ethernet ports using M12X connectors, ensuring vibration-resistant and secure data transmission in applications such as conveyor systems, mobile machinery, or vehicle-based automation platforms.

Application relevance
By combining compact form factor, extensive expansion, and GPU acceleration, the platform enables faster local data processing and reduced system complexity compared to conventional rack-mounted PCs. Typical use cases include visual inspection systems, autonomous mobile robots, industrial IoT gateways, and on-site analytics, where local AI processing can contribute to shorter cycle times and lower error rates without increasing maintenance effort.

All configurations of the platform are available through Acceed, which also provides system integration and long-term availability support for industrial deployments.

www.acceed.com