Industry 4.0 or Smart Industry captures under its umbrella a vast collection of technologies aiming to improve situational awareness, and overall efficiency of operations. At its core are the Internet of Things and Industrial Internet, made possible by edge computing, network technologies such as 5G and 6G, and intelligent automation through artificial intelligence, such as machine learning and computer vision. Edge computing and artificial intelligence make a powerful combination to improve data accuracy, limit latency issues, and create a more realistic big picture of overall operations.
We’re hosting a webinar about AI & Smart Industry on June 24, focusing on Edge AI, 5G, and the practical steps in building distributed intelligence. Check out more information and reserve your spot in the event page.
In this interview, we discuss AI enabling Smart Industry with our Business Development Executive Ville Mickelsson. Ville has 15+ years of experience in working with the leading technology experts from top-tier semiconductors, telcos and IT companies from his role as Nordic Lead for Accenture Industry X.0, and as CEO, first at Sensinode and then later as CEO of CyberLightning.
More accurate data in near real-time
Industry 4.0 lets us gather data from machinery and other equipment at an unprecedented accuracy in near real-time. Its key technology, edge computing, means analyzing and processing the data close to the device or sensors collecting the data.
“With edge computing, we don’t need to store all the data itself or send it to the cloud to get analyzed, but instead, quickly analyze the information that matters. In most cases edge computing also decreases the amount of data transfer needed, resulting in time and money saved.”
Integrating different data sources enables us to build a better picture of the overall situation in the past and right now, thus increasing efficiency and improving optimization.
“New technologies such as machine learning and computer vision enable bringing in more data that further improves the accuracy and predictability of the operations. When combined with edge technologies, we can predict what is likely to happen and with what probability. Better situational awareness and predictability further allows intelligent automation, even in areas with poor connectivity.”, Ville explains.
“Another aspect of critical infrastructure is related to latencies and on what 5G will provide. Not to mention 6G! What will be the physical networks in the future and how do you balance your computing power, network speed, radio penetration, network topology, mix of various frequencies, power restraints of battery equipped devices etc. You can run AI on various devices, concentrators, cell towers, private servers, cloud etc. already today. 5G and 6G will provide partially better answers and solutions for the challenges but not all.”
Optimizing dilemma: lack of big picture
The combination of edge computing and AI makes it possible to create intelligent automation that is able to adapt to changing situations. So how do you start building future-proof operations?
“You’ll need to start bit by bit with different applications, but also make sure that these applications ‘discuss’ with each other. Here it is recommended to avoid vendor lock-ins by investing in open technologies. At Silo.AI, we’ve helped our clients build operations and work flows around visual quality assurance or anomaly detection, predictive maintenance, process quality and efficiency and recommendation engines. These can form strategic steps towards smarter operations.”
“One of the most important dilemmas in optimizing processes here and there is the lack of a big picture. In the worst case, optimizing only parts of the process might even result in negative outcomes.”, Ville says.
For him, organizational culture, internal marketing, brand values and other “secondary” topics might be the boosters, or killers, of building future-proof operations that take big picture into account:
“Your employees, customers, partners, influencers, trends and even weather can also strongly affect your overall operations. Don’t forget them.”
AI for smart industry: computer vision powered cameras and autonomous vehicles
At Silo.AI, we’ve helped various customers further automate their processes with AI. In the industrial setting, our expertise in machine learning and computer vision comes in handy when designing custom solutions and systems that fit into our clients existing operations, thus smoothly improving their efficiency.
“We built an automated visual quality control system for a global leader of renewable solutions in packaging, biomaterials, wooden constructions and paper. In this case, our computer vision based solution learns visual features of each passing flaw, and continuously improves its level of automation and accuracy. The solution improves overall product and process quality, leading to less quality reclamations.”, Ville explains.
Computer vision related AI is to an increasing extent embedded in cameras and other local infrastructure. This enables more rapid and intelligent reactions with minimal latencies.
“Another example is our work with autonomous vehicles, where we’ve developed level 2 and level 3 autonomous driving technologies. In this case, we built a solution leveraging AI sensor fusion for both urban and highway environments. The computer vision powered solution detects moving and static objects in highway situations and detects and tracks pedestrians in urban environments, extending its understanding to the road boundaries, lanes, road markings, signs, and road exits.”
“AI is an excellent tool for building solutions to increase and improve situational awareness, safety, security and efficiency. Similarly to the two cases opened up above, we’ve built awareness modules and systems for challenging industrial settings, such as mines and ports. These solutions aim to track activities and presence of people, vehicles, vessels and equipment. Such environments also depend on edge computing due to poor connectivity.“
Technology agnostic, customizable solutions
The solutions Silo.AI build are technology agnostic and operate with various sets of available sensors, combining vehicles or vessel’s own sensor data with radio signals to produce best overall awareness. In addition, we have built a data labeling tool that uses machine learning to optimize image selection for labeling to maximize productivity.
Ville wants to encourage people to imagine possibilities and to start boldly developing: “All the needed technologies are here already today and there is nothing preventing you from starting your endeavour to empower your business and operations.”
“Think about your customers, partners and the overall ecosystem. What is it that brings all of you the most value? It might be outcome based business model, XaaS, product or just-in-time service that is the key and the right way to provide value. Organization-wide human assisted AI together with motivated employees, robotized processes and engaged consumers is the ultimate goal. The journey has already begun today!”
Interested in building future-proof operations with AI? Get in touch with Ville Mickelsson to learn more on LinkedIn or via email.
On June 24, we’re hosting our next AI in smart industry webinar, focusing on Edge AI, 5G, and the practical steps in building distributed intelligence. We’ll go through a case example of today’s smart harbor, showcasing practical AI-driven improvements for process and operations optimization, supply and logistics chain, situational awareness and overall data streams.
Our webinar will take the form of a round table discussion together with Juha Pankakoski, Executive Vice President, Technologies at the leading crane manufacturer Konecranes, Zach Shelby, CEO and Founder of Edge Impulse, and Ville Mickelsson, Business Development Executive, Silo.AI. Read more and reserve your spot on the webinar page.
