Assembly 4.0 using 5G enabled AI

Dr Krste Pangovski

19 Apr 2022

Krste Pangovski, CEO and co-founder of Vision Intelligence, talks about the company’s involvement in the 5PRING and WM5G programme and the successes of Industry 4.0 use cases in manual assembly.

Krste Pangovski, CEO and co-founder of Vision Intelligence, talks about the company’s involvement in the 5PRING and WM5G programme and the successes of Industry 4.0 use cases in manual assembly.

5G will enable manufacturers to take advantage of the big-data gold-mine, but in the context of exploring the full capabilities of industrial 5G we need to be looking for the gold dust and not rely on finding golden nuggets” – Dr Krste Pangovski, CEO Vision Intelligence

What innovation does Vision Intelligence bring to the market and how does it leverage 5G? The innovation that we bring to market revolves around the application of advanced artificial intelligence in vision systems. Today there are still some 300 million people working in factories, performing manual assembly, contrary to the popular belief that robots have taken over; people are still the dominant force in manufacturing. However, people make mistakes, have different skill levels and or simply have an off-day leading to product defects, variability in product quality and waste. At the same time, we are seeing a significant skill deficit as the previous manufacturing generation leaves for retirement. To answer this challenge we have built VIOLET, an AI system based on vision technologies which aims to support assembly production workers by unlocking the DNA of any manual assembly process thus automatically providing targeted training, assisting the assemblers and providing a quality assurance file for every product manufactured. We use 5G when visual information needs to be available in secure, rapid and low latency environments.

What is your role in the 5PRING accelerator programme? Even the larger companies are still discovering the benefits of 5G. This is the reason why WM5G and 5PRING carries out use cases that are relevant to end users. At Vision Intelligence we bring our existing solutions on the programme to explore how 5G can enable - or significantly enhance - existing solutions or infrastructure. In our case we are exploring the benefits of 5G technology with our existing VIOLET AI solution and other technologies in a manufacturing setting, as part of our 5G enabled Industrial IoT ecosystem.

How was your experience working with Digital Catapult on the programme? It was very pleasant to see the enthusiasm of end users and companies developing 5G technologies. Digital Catapult technologists are able to provide their expertise, so it’s great to have this platform for the use cases.

What business challenges have you been trying to address with the trials at AE Aerospace? AE Aerospace works with aerospace OEMs, with part of the production being in-house and the other part external. We have been working on a technology that detects the movement of a part throughout the production chain and any defects which may arise; should a defect arise, knowing where and why it occurred is paramount. This is essential for compliance, but also for tracing any errors. 5G intervenes in the visual aspect of this defect detection system. With 5G, we can rapidly identify a defect at different points in the production chain, so we can spot the origins of any issues that may arise and ensure they don’t happen again. 5G is very valuable for this use case as it allows low latency and secure data transfer in a private environment, yet make the insights available across the production chain.

How do you view reducing waste and emissions in the context of 5G? If you can’t measure something and extract the data from it, you will never be able to make any real change. To achieve net zero, we need to create a broader sensor network among manufacturers, so we can diagnose all aspects of production. In many cases, a factory has an energy input and a product output. But what happens in between is very complex to monitor, and it gets very hard to identify where the inefficiency lies. 5G is exactly the enabling technology to help us tackle this, given the large sensor base required and volume of data at play. We have an announcement in this space but I can't speak about this just yet!

Were there any unexpected findings from the trials? We thought that current 5G capability would be much less than it actually is. We were pleasantly surprised by the seamless integration of our systems with existing networks. This is very important and powerful because the level of integration complexity is often the failing point for a new technology in a manufacturing setting.

What challenges have you encountered? The challenge ultimately comes down to the use cases. The playing field that has opened up with 5G throws many potential ideas on the table. The difficulty is to sort and pick the best ones. It is a greenfield for innovation and that in and of itself is currently challenging!

What do you see as the biggest barriers for innovators to implementing 5G in a manufacturing environment? There are some technical challenges to overcome. Connecting the technology to the network is the biggest difficulty if native 5G capabilities are to be built into the product; finding technological providers that sells board level 5G components is difficult and since the development of the technology is still underway, the protocols and documentation needs to advance. Further it is important to understand that SME manufacturers are great at what they do, but rarely have the in-house expertise to develop these technologies; this is why agile innovators need to step up to provide turn-key solutions based on a solid understanding of the use cases.

What would be the next step to help SMEs and technology innovators be ready to use 5G capabilities? First, they should look at the challenges encountered by other companies before them. From a technical point of view, they should also look at the availability of 5G sub-components that can be easily integrated into various devices. But most importantly we all need to be open in our early innovation cycle to learn from all use cases, their successes and failures; This is where the likes of WM5G and 5PRING initiatives are of great value.

What are the next steps you will take in terms of 5G? We want to put together a project to enable our edge AI systems to have 5G capability built-in. This will power remote agency and provide remote visual intelligence from anywhere where there is coverage. On the business side, we aim to further develop the use cases, hence the ongoing work with the 5PRING and WM5G networks.

How do you see the future of Industrial 5G? What 5G enables is breakthrough in use cases and products, but how we get there won’t be a smooth, steady road. Rolling out 5G technologies across the country will require collaboration between network operators and end users but this should not run too far ahead of understanding the end use cases; we could end up in a situation where the network architectures do not support the most pertinent use cases. I also think that the future of 5G will in large part be determined by innovative SMEs - who, due to their agility - can be the driving force pushing the use cases and technology roadmap forward; This is what I refer to as the gold dust.