Application Highlight: AI Edge Inference with Premio RCO-6000-CML

March 11, 2024

Sponsored Story

The place where Edge Computing meets AI is where every company wants to be, if it’s trying to get a piece of the explosive growth occurring thanks to the Fourth Industrial Revolution. And square in the center of that critical intersection is the Inference computer.

In a system that incorporates sensors and processing at the Edge, and layers AI algorithmic intelligence over that data, a key goal has always been for the outcome to be an embedded system that can use those environmental and systemic data inputs and infer appropriate actions to be taken under dynamic conditions.

One company has been doing just that on its in-line bottle inspection facility.

Application Use Case

A major in-line bottle inspection solution provider has been bringing its solution to the rescue of major F&B companies all over the world for more than 60 years. The company has always prided itself on offering automated state-of-the-art bottle quality control systems, and its computer vision capabilities give it the edge thanks to being integrated onto high-speed production lines and ensuring the examination and analysis of each and every product that rolls off the line and heads out of facilities to the shelves of wholesale-retailers.

This company’s in-house interface uses high-fidelity machine vision to inspect for empty bottles or cans, pressure closures, integrity of tamper bands, proper fill levels, and many more critical parameters and KPIs. Having this inspection process automated gives the food and beverage companies the confidence that their products are meeting quality standards, and the data to prove it, which reduces the risk of delivering faulty products and maintains brand reputation.

With all of these advantages in mind, and operations rolling successfully, the company wanted to enhance its offerings with a reliable embedded edge computing solution that would help to streamline its machine vision systems. Previous attempts to find the right partner resulted in poor communications about design layout changes without prior notification and other problems, so any new partner had to bring a superior level of professionalism and competence.

More than just adding Edge Processing to the inspection system, the company also needed to upgrade its connectivity from an old model Intel 7th Gen session border controller (SBC) to a more modern, ruggedized industrial embedded computer in order to qualify for EMC certifications. It also wanted to improve efficiency by consolidating disparate system solutions into one computing platform that would save money and reduce form factor.

Facing all of these challenges, while keeping production lines at full capacity and avoiding supply chain problems meant a local and well-connected partner was critical to success, in addition to finding a provider with top end technical chops.

Premio RCO-6000-CML Series AI Edge Inference Computer

It didn’t take long for the company to settle on Premio’s AI Edge Inference Computer, the RCO-6000-CML Series. First of all, Premio’s location in California offered the localized supply chain management and local, dedicated engineering support that the customer wanted. Even more critical is that the RCO-6000-CML series of embedded computing solutions offers high-performance processing and flexible connectivity that’s specifically designed and engineered for machine vision applications that require rugged edge compute.

The on-board Intel XEON-W TE processor brought server-grade performance and reliability at the edge, with an extended product lifecycle roadmap, and made the entire system eligible for the required EMC certifications, too. Also bringing the inspections systems into regulatory compliance, the fanless and cableless designs have already been evaluated and validated for safety standards under CE, FCC, UL, and more.

The RCO-6000-CML also consolidated legacy systems into more efficient designs by increasing simultaneous camera-to-module rations to 8:1 using powerful EDGEBoost I/O modules. One key differentiator of the series is the two-piece modular design, which creates flexibility for system integrators that want to tune systems to workloads by choosing different performance-based nodes. Users can select performance nodes that provide options in NVMe storage, high-density SATA storage, and even GPU/M.2 acceleration modules for real-time machine learning applications.

Thanks to the dual slot bracket for modular I/O daughterboards, the RCO-6000-CML series supports up to x8 additional LAN & PoE in wired RJ45/M12 connectors, x8 USB 3.1 gen 2 ports, x4 10GbE in RJ45 connectors, and a 5G-ready module for low-latency wireless connectivity.

Check out a video of the system in operation here.

With key concerns for edge processing and low latency, Premio was able to give the customer top performance with the Intel 10th Generation CML S Processors and W480E Chipset support, along with the LGA1200 socket design and Intel W480E chipset to deliver augmented peripheral performance for low latency edge responsiveness. If it’s required for a given application, customers can also choose to use Intel XEON processors for server-grade performance in a fanless thermal profile.

The XEON W-1290TE is a 35W TDP processor that delivers 10 cores for multitasking and even supports error correction code (ECC) memory for data redundancy in mission-critical applications. The XEON processors make any AI-powered machine vision system meet the most demanding reliability and performance benchmarks, and lets the system support gigabit wireless speeds, PCIe 3.0 lanes, SATA ports, and high-speed USB 3.2 Gen 2 for transmitting data to and from sensory devices sitting at the edge.

In terms of environmental resiliency, the design of the RCO-6000-CML as a fanless, rugged computing solution makes it capable of powerful compute at the edge even under harsh environmental conditions. This includes temperature ranges of -25ºC to 70ºC, input voltage spreads from 9 to 48VDC, and even resistance to shock at up to 50G and vibrations at up to 5GRMS.

Premio has brought together a wide variety of optimized technologies to create a real-time responsive embedded computing solution at the edge, layered with the processing for AI intelligence that’s needed to receive and offload volumes of rich data. IoT integrators and industrial automation operators can rely on the RCO-6000-CML AI Edge inference computer to manage the most complex workloads in space-constrained deployments in the harshest environmental conditions.

It's worth a closer look at this machine vision evolution.

Additional Resources:

• Product Page: https://premioinc.com/pages/ai-edge-inference-computer-with-10th-gen-intel-core
• Product Brief: https://premio.blob.core.windows.net/premio/uploads/resource/product-brief/PB_AI%20Edge%20Inference%20Computer_RCO-6000-CML.pdf
• Data Sheet: https://premio.blob.core.windows.net/premio/uploads/resource/datasheet/RCO-6000-CML/DS_RCO-6000-CML_Premio.pdf