Intel plans to boost its presence in the IIoT segment with Edge, Deep Learning-based Computer Vision, Open Architecture, and bringing developers together.
The role of Industrial IoT (IIoT) is all set to become more and more critical in the coming years, as companies think outside the box to boost efficiency and cut costs. Intelligent sensors and networked devices would have a critical role to play on factory floors that have, over the recent years, faced concerns like talent shortage.
But
IIoT is still at its nascent stages and continues to evolve as customers and solution providers attempt to tailor it to their requirements. Several challenges persist, and solving them is a critical part of the IIoT adoption story.
A key player in this field, with a keen eye on solving customer challenges, is Intel. The company has optimized its chipsets over the years to enable more and more AI workloads to take advantage of this. Recently, at the Intel Industrial IoT Summit 2020, the company launched its latest 11
th Gen Intel Core processor that harnesses the most advanced technology to make factory floors future-ready by enabling them to draw timely insights from data that sensors collect.
Speaking to asmag.com on the summit’s sidelines, Eric Chan, VP and GM of IOTG Customer Engineering at Intel, elaborated on this, explaining how the company goes out understanding customer pain points and resolving them.
“The interesting part of my job is understanding customer pain points and how we can deliver the solutions to solve them,” Chan explained. “It will take a while for the technologies to mature, but we are going to great lengths to make the process easier for our clients.”
Driving more powerful edge solutions
We are witnessing an exponential increase in data generated. Analysts expect that by 2023 more than 50 Terabytes of this data generated in various industries, including factories, would be processed at the edge. Processing at the edge removes many of the challenges that customers face concerning latency, bandwidth, reliability, security, and privacy.
Intel, on its part, is doubling down to cater to the demand for silicon chips this growth in edge solutions would bring about. Notably, the 11th Gen Intel Core processors, Intel Atom x6000E series, and Intel Pentium and Celeron N and J series processors revealed at its Industrial summit respond to challenges in edge complexity, the total cost of ownership, and environmental conditions.
Encourage openness for better AI training
On the silicon side,
the systems must be AI-ready. This means that solutions should know what to look for right from the moment it’s deployed. Chan explained that for this, Intel’s OpenVINO toolkit helps customers optimize the workload at different points – whether at the edge or on the server.
“We support cloud-to-edge systems: not everything is processed at the edge,” Chan said. “In some cases, you need to go back to the cloud. The OpenNESS (Open Network Edge Services Software) toolkit and Intel® Edge Insights for Industrial software allow us to seamlessly integrate devices and bring the cloud’s power to the edge. Industrial customers need to combine different systems. But ingesting data is not easy; it requires training and then inferencing. To simplify this process, Intel allows customers to easily train the system with the data they have.”
Leverage Deep Learning for better computer vision
Although more and more factories make use of Computer Vision (CV) to automate processes, challenges like difficulties in scene and background recognition persist. Chan acknowledged this and explained how Intel approaches such problems.
“Until recently, CV was all about pixel manipulation,” Chan said. “But that is a traditional approach with limited efficiency. In some cases, you can get good results, depending on light and resolution. You may even get 90 percent accuracy. But the future is about using deep learning (DL). To pull it off, you need a big data set (for example, for LPR). Only through DL training can you get to a point when you have close to 100 percent accuracy. The Holy Grail is high accuracy with a small data set and short training. But you need to have a few good data sets to get good results.”
To clarify his point further, Chan cited a case study. A welding company needed assistance in defect detection. The management was using traditional CV for this, but the best they could achieve was 90 percent accuracy. Intel managed to ramp this up to 95 percent with AI, which is a significant improvement at the highest levels. In a few years, machine vision would develop further to provide 100 percent accuracy and even replace humans in certain scenarios like harsh environments.
Connecting developers to customers for innovative solutions
An important point to note is that technology is increasingly democratizing many industries by giving more people access to contribute. Intel understands that there are several skilled developers worldwide, and orienting them to the right market could accelerate developments in IIoT. This is yet another step to solve customer pain points, as well.
“The traditional ODM and OEMs that we work with are good, and they continue to do their part, but a lot of the AI applications are coming from the developers,” Chan said. “These developers may or may not have an industry focus. But we believe that it’s a good idea to work with these AI developers to provide them access to the industry while helping our customers make use of their innovations. So, we also have an outreach program for the developers.”
How systems integrators benefit
The benefits to customers are clear, but how does Intel aim to work with systems integrators (SI)? Chan explained that some customers prefer to have a third-party to help them set up and maintain their systems. In such cases, the company can work through an SI to ensure that the end customer receives maximum benefit.
“We make sure that these SIs are provided the best possible resources that they need to complete their projects and continue to serve the customers,” Chan added.
In short, Intel’s ability to think ahead and come up with solutions for the future, along with its sensitivity to customer requirements, is what makes the company an undisputed leader in IIoT. Paying close attention to every challenge that the customer faces is a priority for the company. Combining it with a culture of innovation makes it a success story.
Click here to learn more about Intel® Edge Insights for Industrial