The Rise of Artificial Intelligence Machine Learning in Manufacturing on the Factory Floor…
By Ashley Ziatopolsky, from the AME 2024 Program
In the world of manufacturing, a new application and subset of artificial intelligence, or AI, is allowing machines to learn from data without needing to program them to do so.
It’s a growing technology known as AI machine learning, or the ability to simulate human thinking capability and behavior, and it’s poised to shake up the manufacturing industry for good.
Mark Ermatinger, CEO of Industrial Control and founder of the Advanced Manufacturing Expo, shares why AI machine learning is here to stay.
Attend AME 2024! Get more information here
How has AI already transformed the manufacturing industry as a whole?
The rise of AI is changing how we check for quality issues, assist assembly and help with safety on the factory floor. In manufacturing, AI is mainly used for quality control on machine vision cameras which traditional rules-based vision systems wouldn’t work well on. Yet more and more, it’s now being used to watch operators’ assembly parts and alert them if they skip steps or insert components the wrong way. For example, AI will throw a light down on a machine operator’s surface if something is done incorrectly.
AI is also excellent for reading text, especially a technology called Optical Character Recognition (OCR), along with identifying what part it sees, such as colors and wood grains. As AI is quickly emerging, it’s our job at Industrial Control to vet and test different technologies to figure out if they will hold up within our factory.
What types of AI machine learning has Industrial Control deployed so far?
Our company is currently deploying fi ve different AI machine vision projects from four different providers. One project uses AI to inspect taser parts to ensure they’re made correctly before they’re shipped to their customer. Another type of AI machine learning we use checks for defects in the paint on automotive parts, so things like dirt spots and tiny scratches [that humans may not see as easily].
Now, we have seven AI vendors compared to three last year. We’re also talking with another vendor that works with audio AI, where we can “listen” to machines to predict machine failures before they happen. That one is still in development. So, let’s say you have something on the machine that squeals when it’s too hot, you can teach AI that a temperature setting
is too high, and that sound happens when there is a problem.
We’ve also helped local manufacturers deploy autonomous floor robots that can reroute themselves as needed when the aisle is blocked to reach their destination, but with true machine learning, these same robots will learn which aisle is typically not blocked as much and will choose that route as the default - another AI capability.
Does AI machine learning pose any challenges to manufacturers?
The number of new companies offering AI for manufacturing is astonishing, especially in the last few years. Yet when you bring these new technologies on the factory fl oor, you may have a whole set of issues. The biggest limiting factor is the computing power and the heat AI technology produces. AI requires very powerful multicore central processing units and graphical processing units. AI servers can’t tolerate heat well. If it gets too hot, they start to degrade their performance, so you’ll need proper air conditioning or heat dissipation to maintain their quality.
What role do you see AI machine learning playing in the future of manufacturing?
AI is on the move in every aspect of our lives, including the migration into industrial robots that many manufacturers use to produce products and goods every day. Currently, there is a lot of work being done with Natural Language Processing (NLP), which is a branch of artificial intelligence for programming industrial robots to understand text and speech much as a human would. One example of a text string is “pick up the part off the conveyor, place it in the fi xture and then apply the glue bead.”
I think we are a few years away from this level of understanding for robots, but brace yourself and lean into it, as these new tools will accelerate Industry 4.0 in a whole new way and likely in every aspect of your company. There’s also a lot of focus right now on setting up machines to run even faster and for robots to do what they need to do without programming a lot of code in a special language. Once we get to that point, we can just type in the words that we want the robot to do, and then they’ll be deployed much faster.
The market for AI robotics is also growing exponentially. Marketsandmarkets.com tracks these changes and reports that the AI robotics market stood at U.S. $6.9 billion in 2021. Now, it’s forecasted to reach a staggering U.S. $35.5 billion by 2026 at a compound annual growth rate of 38.6%, meaning AI machine learning is definitely here to stay.
Is AI machine learning pushing manufacturers to reinvent their digital strategy?
I think so. It’s not just manufacturers, but anyone in any type of work. People are moving forward right now by experimenting with AI in all aspects of business. For us personally at Industrial Control, we’re going to need to look at every aspect of every scale. We’re going to use AI so that we can be the disruptors going forward.
To keep up with these changes, it’s important to understand exactly what AI machine learning is and how it’s one level above your traditional, everyday AI. AI machine learning is a subset of AI that allows machines to learn from experience and improve over time. AI can perform tasks like seeing, understanding, responding to language, analyzing data and making recommendations. AI machine learning, meanwhile, can perform even greater tasks like automatically learning and improving, identifying patterns and performing specific tasks - making it the technology of the future.
Interested in learning more about our ai solutions?
Back to All News