HMIs increasingly have to present a visualization of data and decisions made by edge AI devices and products. The solution is more intelligent displays, according to ADLINK Technology and AUO.
There are two adjectives that we see used a lot these days to describe almost everything: ‘smart’, and ‘intelligent’. Whenever I see these used to describe any new product or technology, I often ask myself: “What do they actually mean? Is it simply a device with a bit of code embedded to enable simple actions on certain triggers from sensors, or does it actually have some level of intelligence built in to actually make decisions based on the sensor input?”
In the case of the latter, one area that has emerged in this new age of intelligence at the edge is the intelligent human machine interface (HMI), which increasingly has to present a visualization of data and decisions made by edge artificial intelligence (AI) devices and products.
Last year, two companies that represent the two elements of this capability came together to offer intelligent displays. We talked to the two companies, ADLINK Technology and AUO, to understand some of the areas they are working on for what they call edge visualization solutions, or intelligent displays. AUO specializes in high end displays, with considerable success in the automotive industry, and ADLINK is focused on embedded computing solutions offering edge intelligence.
The two companies said their partnership eliminates the need for customers to source the display and the control electronics separately, as they can then offer an integrated solution or subsystem for enabling intelligent displays.
Jim Liu, CEO of ADLINK Technology, explained, “There is a tremendous synergy from our two companies working together. AUO’s business model is to provide high end panel displays that meet the specific needs of various application areas. For example, easily cleaned to the high standards needed in an operating theater or with the long operational life required by the automotive industry. Similarly, ADLINK tailors computing solutions for applications from factories to railways where lives could be placed at risk if the equipment does not perform correctly all of the time hence our factories operate to the highest quality certifications.”
He said that ADLINK had already been working with AUO for over 10 years to create products that use AUO displays for the its medical display division, and that alone had a revenue of over €23 million (approximately US$ 25 million) last year. The key products are intelligent displays that are used in operating theaters for key hole surgery, for example, where the surgeon needs high resolution images and video that can be 4K and even 8K to provide the level of detail needed.
“We developed a close working relationship with AUO from doing this,” he added. “We explain to them the new trends that we see coming up for displays so that they can develop solutions and, conversely, they tell us what intelligent features their customers want built into their displays. We have now formalized this into a joint service for customers that want specifically designed, intelligent displays that are made to mission critical standards.”
Frank Ko, president and COO of AUO, added, “We have huge experience in manufacturing displays to meet the demanding specifications of mission critical applications. In almost all of these applications, the HMI is absolutely vital as so much information can be displayed quickly via a display.”
He said the displays have to be extremely robust for these applications and are put into enclosures laminated with protective glass to protect against impact and sealed against dust and liquids to IP54 ingress protection level. The latter is particularly important in industrial and medical applications to ensure that they can be vigorously cleaned for the lifetime of the display, which for such high value equipment can costs thousands and be expected to last 10 years or more. This long lifetime expectancy from customers means that the displays have to have even longer product availability so that spares are available.
“There is a world of difference in making these displays and making TVs or computer monitors,” added Ko. “Each of our products has a wealth of detailed information from every stage of the testing so that in the highly unlikely event of an issue it can be traced to the cause. If it was not a one off, then pre-emptive, corrective measures can be done for similar devices rather like the very rare case of a recall for a model of car. This level of detailed audit is second nature to us as it is to ADLINK, which is why together we form a partnership that companies can rely on to deliver intelligent, start-of-the-art displays with the highest levels of reliability possible.”
AUO said it has been granted more than 21,700 patents globally, 97% of which are invention patents. “Whatever bespoke specification a customer needs, we will have a solution in our patent portfolio,” added Ko.
The intelligent display capability builds on the working partnership that ADLINK and AUO have developed when working on another area, edge computing for industry 4.0. When raw data is gathered at the edge, it needs to be processed and presented to operators so that they can action decisions. The companies said this is best done visually by pairing AUO displays with analytics from ADLINK to create ‘edge visualization’.
This elevates the HMI from just displaying information to displaying knowledge; in other words, not just what is happening but why it is happening using AI software to interpret events so that informed decisions can be made faster. For example, an event only happens when two separate parameters are exceeded. This also helps with predictive maintenance, for example, detecting very early when something is starting to go wrong so that corrective action can be taken before expensive, unscheduled down time is needed for repairs.
Speed and image quality are also necessary for edge visualization to be useful. For example, in a medical application such as an endoscope image it needs to be virtually real time (less than 50 milliseconds latency) otherwise the tip of the endoscopy could damage tissue. ADLINK has developed a solution that doubles the data transmission throughput for better images and almost halves the latency. Traditionally, incoming data goes into the main memory via the CPU and then out to GPU for display. ADLINK said its new method leverages remote direct memory access (RDMA) to bypass the CPU stage so that the incoming data goes directly to the GPU to save time. In addition, ADLINK uses the parallel processing capability of an Nvidia CUDA GPU which is like having hundreds of CPUs working simultaneously on the data to create more detailed images faster.
Another example is in railway applications with ADLINK’s AI-enabled video analytics platform that again uses an Nvidia GPU for video/graphic processing and forms the heart of monitoring and safety systems for increasingly complex railway operations that need real time information. The units are invariably located in remote places alongside tracks or on trains so that have to be very rugged to withstand extremes of weather and vibration if on a train. They perform mission critical functions such as railroad intrusion detection, hazards on the line and station video surveillance where lives could be at risk.
This article was originally published on Embedded.
Nitin Dahad is a correspondent for EE Times, EE Times Europe and also Editor-in-Chief of embedded.com. With 35 years in the electronics industry, he’s had many different roles: from engineer to journalist, and from entrepreneur to startup mentor and government advisor. He was part of the startup team that launched 32-bit microprocessor company ARC International in the US in the late 1990s and took it public, and co-founder of The Chilli, which influenced much of the tech startup scene in the early 2000s. He’s also worked with many of the big names—including National Semiconductor, GEC Plessey Semiconductors, Dialog Semiconductor and Marconi Instruments.