How next-generation technologies are driving Industry 4.0
Posted by Ido Gur on Jun 15, 2020
Tags: • 5G • IoT • edge computing • Industry 4.0•
Empowering the fourth industrial revolution with smart communication technologies
The fourth industrial revolution is underway, and it’s transforming the world of production and manufacturing. Lights-out manufacturing, in which factories can function with little or no human supervision, is helping industry become safer and more sustainable. Connected systems on production floors are being used to optimize industry management. Sensors are being used to monitor environmental conditions and industrial control systems for predictive maintenance and risk management.
These are just some of the ways cyber-physical systems are changing the world.
Bringing computing infrastructure to the edge
IDC estimates that, by 2023, more than 50% of new computing infrastructure will be deployed in edge locations rather than corporate data centers. Today, the most common approach is to transmit data to a centralized system, such as a cloud platform, for processing. Naturally, this introduces some unavoidable limitations with regards to bandwidth and latency. Often, data has to pass through multiple physical systems spread across a large geographical area. With edge computing, data is collected, processed, and analyzed on a local device supported by a local network.
Traditionally, data analysis takes place externally, which ultimately delays decision-making. In other words, it takes more time to identify and resolve issues. Plant managers often have to wait for lengthy analyses before they can act. In many routine processes, the delays might not present a serious problem, but in other cases, any unnecessary delay can lead to disastrous consequences. For example, if a machine on the factory floor starts to overheat, it’s not much good receiving a notification after the damage has already become irreversible. But, with edge computing, it’s possible to automatically adjust processes to prevent damage and ensure the safety of workers on the factory floor. This reduces operating costs and risk alike.
Facilitating millisecond response times with 5G
When it comes to industrial control systems, response times are everything. Transferring data from sensors on the factory floor over to a cloud data center, which could be hundreds of miles away, takes time. Moreover, conventional wireless networks don’t offer the coverage required in larger production shops without using range extenders. These increase latency and reduce performance anyway. Previous generation 4G mobile networks aren’t an adequate solution either, since they need to transmit data via an external network operator, which also leads to reduced speeds and longer delays.
5G changes that with the introduction of private mobile networks. In essence, these are the next generation of local area networking. Bringing latency down to just a few milliseconds and enormous bandwidth to a huge number of devices, private 5G networks are a key driver in the adoption of edge computing networks. As a result, it’s possible to build ultra-fast, ultra-reliable local communication networks that are ready to accommodate the needs of complex industrial IoT infrastructures.
Closing the control loop with artificial intelligence
Going back to our previous example, when a machine overheats, it will either be necessary to turn it off or change its operational variables or surrounding conditions so that it can continue to operate safely. Even if plant managers do receive the alert in time, they still need to act. They might only have time to sound the alarm and get everyone to safety if there’s a risk of fire or explosion. In the meantime, there may be a high risk of serious and costly damage or worse, personal injury. But no matter how astute the plant manager might be, humans are still slower than computers.
Artificial intelligence closes the loop by taking the sensory information from IoT devices on the factory floor and acting on it in real time. But AI is also a data-heavy and computing-intensive, hence the need for edge computing to process, analyze, and act upon that information in real time. One of the best ways to achieve this is through GPU virtualization, which uses cutting-edge performance hardware for real-time AI processing right there on the factory floor.