NEXCOM helps cities go farther with AI on the edge

Gateways with Coral intelligence can analyze and deliver key information exactly when and where it’s needed.

ic_share_24px Created with Sketch.
Close Created with Sketch.



The data that highway departments need to effectively plan roadwork; factories need to maximize productivity; and industrial plants of all kinds need to head off unplanned shutdowns often comes too late. But this could change with new sources of information about the vehicles and vehicle types on the road at any given moment; about how temperature variations affect the quality of baked goods in real-time; and even how subtle changes in vibrations can predict impending equipment failure.

To help solve the problem, NEXCOM—an industrial computing company with two decades of experience—is using Coral products built into Internet of Things (IoT) gateways that analyze and deliver key information exactly when and where it’s needed.

We provide edge AI compute capability to independent software partners.
Alexander Su
General Manager, NEXCOM, USA

Next-level IoT

NEXCOM’s IoT gateways with Coral intelligence act as ML-enabled connecting points or pathways for data flowing into or out of systems ranging from smart traffic cameras to factory machines.

Alexander Su, general manager of NEXCOM USA, explains that his company designs hardware that makes a wide range of use cases possible: “We provide edge AI compute capability to independent software partners.” NEXCOM's partners, in turn, provide IoT solutions to factories, city governments, and other customers.

With the Coral board, they were further ahead than anybody else.
Tom Arthur
CEO of IoT at Dianomic

Connectivity for AI

“The most important thing for these gateways is the connectivity,” Su says. That includes ports for connecting sensors such as cameras, displays, and other peripherals, as well as Ethernet, Wi-Fi, or cellular connectivity to send data on to other systems.

Now, thanks to Coral Mini PCIe Accelerators, each NEXCOM gateway can process data locally, using less power, and without having to send data to the cloud. This means that critical systems can act more quickly, with less latency and more privacy.

“Coral powers high-performance applications on the edge by enabling quick and easy TensorFlow Lite models to be deployed," says Daniel Man, a software application engineer with NEXCOM.

In other words, Coral enables machine learning (ML) models to run where they’re needed, faster, saving precious seconds in the field. Seconds in which factory machines can keep up with production lines, traffic can get counted in real-time, and plant managers can predict when maintenance will be necessary and plan it for the least disruptive times.

Keeping it local for AI processing also enhances security. “They don't want to share their factory data,” Su says of plant managers. “They want to keep their data internally and analyze their data internally.”

Nexcom’s Coral-powered gateways include AI Edge Computers and AI Edge Telematics Solutions for use cases such as smart signage in retail applications; for manufacturing applications; and for smart traffic control.

The future of automation

Data security isn’t just a good idea; it's a regulatory requirement in critical infrastructure such as power plants. “In many industries, data never goes north of the firewall,” Tom Arthur, CEO of IoT software developer Dianomic says. “All of that operational data is staying south of everything. That is generally all about security.”

That’s why Dianomic deploys NEXCOM gateways to give industrial customers anomaly detection, monitoring, and control capabilities that don’t require sending data to the cloud. Arthur says the out-of-the-box solution that his company’s FogLamp software provides is a rarity in the industrial space, saving customers money and time without requiring specialized expertise. “I’m using NEXCOM from A to Z,” Arthur says, referring to the range of capabilities his customers need. “With the Coral board, they were further ahead than anybody else.”

Arthur says his customers are mostly still adding the most basic monitoring functions. But he foresees demand for local AI capabilities, which is why Dianomic is developing applications for them. One example: safety alerts that warn workers who stray too close to dangerous machinery or forget to wear hardhats.

“Those kinds of applications are pretty universal,” Arthur says. And Dianomic is doing its part to get them to market, with the help of NEXCOM gateways powered by Coral's AI solutions.