by Mika Skarp
Automation in factories has developed in phases from the early days of industrialisation. In the beginning it was about having more production and lower prices. In the second phase tasks that were dangerous or very difficult for humans to do were automated. In some cases, automation has been mandatory because of production environmental requirements; a good example of this is the need for clean rooms when manufacturing semiconductors.
The next phase of automation in western countries is driven by productivity. The population in western countries is getting older, meaning societies need to be organized differently. Work that can be automated has to be automated, allowing the workforce to be used on areas that are more difficult to automate. In practice this means that factory automation will allow human labour to be used in hospitals, schools etc. Today we can see two different kinds of trends in automation. Machines need to be able to make decisions by themselves and they need to be able to move around. In technical terms we need AI and mobile networks.
How will the factory of the future look? How can we describe the software-defined factory? Answering that question is not possible; let me explain why.
Limits of the Future Gaze
At the end of 1980s telephones were connected to the wall using copper cable. The phone was in a fixed position. There was some kind of mobility available using many different technologies but for common people the phone was fixed. By phone you were able to order a pizza, ask the time and do many other things just by calling some other human being.
Back then, if you were asked if you believed someday everybody would have a mobile phone, most of us would have said “yes, it sounds possible – maybe not revolutionary but feasible.” It was something that you were able to easily imagine.
At the same time there was a data network which allowed computers to communicate with each other around the world, used for example by banks and universities. If you were asked if you believed that someday everybody will have a computer that can communicate with other computers worldwide, most of us would have said yes, that is possible.
Now if you were asked if the same device could be both a mobile telephone and a computer at the same time, most of us would have been really confused. We would think it made no sense to have such a device, that it would be so expensive and bulky that it would be impractical. We would not have been able to understand how two new things would have been used together.
How would you describe a mobile phone application to a person living at the end of 1980s – you can order a pizza when driving a car? A music service would be easier to understand, but something like TikTok would be more difficult to explain. Understanding the possible business models around these innovations would be a huge obstacle to envisioning the services.
Today we are in the same situation with AI and mobile technology in factories. We can understand that AI will be very useful in factories, and we can understand that machines without cables would also be very handy, but we are not able to describe if and how these two new technologies can work together.
So we must build the future ourselves. We will move forward step by step, understanding that we’ll make some mistakes and most likely we’ll see hype and stock bubbles. Changes will take more time than we anticipated, and the change will be more dramatic than expected. Changes are not happening overnight but slowly and almost invisibly.
AI is coming into factory offices right now in the form of ChatGPT language models and soon we will see production plans and daily product changes planned by AI to optimise energy consumption, waste or other factors. Machines will use non-public wireless networks to connect and some machines will have batteries as an energy source. With mobile machines we can make factory floorplan changes faster, we can move products in the factories without human touch, we can integrate the warehouse with the factory and so on. But for sure there is much more that we don’t see at the moment.
6G will turn mobile networks into data sources. Mobile networks can start sensing the environment. Location information has been available for some time, but velocity can be measured and, in the future, also temperature and position – to identify, for example, if a person has fallen. There are many things that we can use radio signals to measure; just think about person and luggage scans at the airport.
To make machines mobile we need to allow them to communicate with the existing infrastructure today. Factory automation happens on Layer 2 and mobile networks are working on Layer 3. The advantage of using cellular technology to allow mobility of machines is huge. Mobile networks have been designed to work in the environment with many regulated frequencies and can thus be used on higher output power than proprietary networks or Wi-Fi networks. Mobile network devices are manufactured in large volumes so there is an economy of scale. Mobile networks also have mobility built into the design; session continuity is there, so the only thing missing is the Layer 2 communication.
To address this problem 3GPP has developed a feature called 5GLAN. By using an Ethernet protocol data unit (PDU) instead of regular PDU we can grant factory automation devices mobility using the regular mobile network. Cumucore has worked together with partners to showcase 5GLAN first time in Europe.
After 5GLAN is in place, we can synchronise these devices. This functionality is called Time Sensitive Networking (TSN), and from there move on to other services. As mentioned, we will get to the future factories step by step. 5GLAN is the first step, just as a text message service was the first step towards the mobile internet.
Long time reader, first time contributor. Love technology and the great outdoors. Looking forward to discussing everything beyond 5G and the future of wireless technology!