Transitioning Embedded Prototypes to Industrial Edge AI

SECO releases the SOM-SMARC-Dragonwing-IQ8, a hardware evaluation platform developed alongside Arduino and Qualcomm Technologies to support the deployment of industrial Edge AI applications. This module allows original equipment manufacturers (OEMs) to transition AI-enabled robotics, industrial vision, human-machine interfaces, and machine control architectures from rapid prototyping environments into production-ready deployments.

Bridging Prototyping and Production Architectures
The integration between the Arduino VENTUNO Q development platform and the new industrial module addresses a common gap in hardware engineering. While prototyping platforms allow engineers to validate initial concepts quickly, industrial deployments require rigorous hardware standards for reliability, cybersecurity, and long-term lifecycle management. Developers can build applications within the prototyping environment and port them directly to the industrial system-on-module using software tools like Arduino App Lab.

Representatives from Arduino and Qualcomm Technologies highlighted that this continuity preserves the initial software codebase and validates the system for immediate integration into production-oriented architectures. The process relies on a common hardware and software foundation to reduce development effort and accelerate the time to market for compute-intensive devices.

Hardware Specifications and Fleet Management Integration
Built according to the SMARC Release 2.1.1 standard, the module utilizes the Qualcomm Dragonwing IQ-8275 processor to deliver up to 40 Tera Operations Per Second (TOPS) for AI acceleration. The hardware supports high-performance operations through configurations offering up to 32GB of LPDDR5 or LPDDR5X memory alongside a maximum of 1TB of UFS 3.1 storage.

To accommodate industrial environments, the module features extensive connectivity options including Gigabit and 2.5 Gigabit Ethernet, PCIe Gen4, MIPI-CSI, and CAN-FD interfaces. Software operations are managed through Clea OS, which provides secure over-the-air (OTA) updates, container orchestration, fleet management, and remote troubleshooting to maintain device scalability across distributed networks.

Additional Context
This section details technical specifications and competitive benchmarking not included in the original news release.

In the industrial Edge AI module market, the 40 TOPS processing capability of the Qualcomm IQ-8275 processor places it in direct performance competition with the NVIDIA Jetson Orin Nano module, which also delivers 40 TOPS for localized artificial intelligence inference. While the Jetson Orin Nano derives its computational power from an integrated Ampere architecture graphics processing unit (GPU), the Snapdragon-based Dragonwing architecture utilizes a dedicated neural processing unit (NPU) to manage artificial intelligence workloads with optimized power efficiency.

Furthermore, the hardware adheres to the vendor-agnostic SMARC 2.1.1 physical footprint standard. This compliance allows hardware engineers to interchange processing modules on standard carrier boards without redesigning the underlying physical printed circuit board infrastructure. In contrast, the Jetson ecosystem utilizes a proprietary 260-pin SO-DIMM physical form factor, which restricts engineers to specialized carrier board designs dedicated exclusively to NVIDIA silicon architectures.

Edited by Aishwarya Mambet, Induportals Editor, with AI assistance.

www.seco.com