The speed at which technology is advancing can seem disorienting as we all try to keep up.
And while it may not be so apparent now, these rapid advances are helping make complex
technologies much easier to use.
In a featured meeting at Automation Fair 2024 between Blake Moret, chairman and CEO of
Rockwell Automation, and Rev Lebaredian, vice president of Omniverse and simulation
technology at Nvidia, the two discussed how simplification is arising from complexity.
“What were experiencing now with AI (artificial intelligence) may seem discombobulating,
but it’s quite simple,” said Lebaredian. “A little more than a decade ago, AI and machine
learning introduced the possibility of solving computing problems that we couldn’t resolve
before. Things like image classification and computer vision that we had failed to develop in
a robust way. But now, with AI software that can write the algorithms simply by feeding it
examples of what we want, we can now create these advanced systems with ease.”
Moret added that Rockwell Automation is focusing on AI for its simplification possibilities.
“A lot of the things we’re applying AI to are to simplify the whole business of designing
systems to be able to commission them with simulation as well as operate and maintain
them in a predictive way. I can remember old use cases related to machine vision where it
was just too difficult to hard code the classification systems for sorting, or to look for
imperfections on a metal surface. AI gives us the opportunity to be able to apply
sophisticated sensor data at speed on a line or in a piece of equipment.”
Another example of the simplification AI is bringing to industry, Moret noted, is by using AI
copilots that allow an operator or engineer to use their natural language to program a Logix
controller.
“People ask if this is a future capability,” said Moret, “but it’s here today and it’s going to
allow people to not worry about the arcane syntax of ladder logic if they prefer to program
using natural language.”
Avoiding waste with simulation
Lebaredian stressed why Nvidia “deeply believes in the power of simulation.” He said it’s
because Nvidia uses it to build its chips so that they can be certain the chips will work.
“Before we send a chip file to be fabricated, we have simulated every possible
configuration and emulated the running of applications on it so that we’re certain that,
when the chip comes back, it’s going to work,” he said.
He added that Nvidia believes this same approach is necessary for all things built in the
real world, especially as they get more complex with built-in technologies.
“Simulation is the only way to get good time-to-market and not be wasteful with materials
and energy,” said Lebaredian. “With Omniverse, we’re devising ways to build large scale
simulations and integrating that into Rockwell Automation tech like Emulate3D because
it’s not just nice to have that capability, it’s essential to get to the next wave of the industrial
automation revolution, which involves adding intelligence to complex systems.”
Moret gave an example of how AI-driven simulations are making complex commissioning
applications possible. “Consider the commissioning of a bottling line,” he said. “You
couldn’t have people representing each piece of equipment — like the depalletizer,
conveyor, labeler and bottle washer — from all the different suppliers on site at the same
time. It just isn’t possible to bring everybody shoulder to shoulder in a normal
commissioning process. So, the idea of being able to remotely commission with digital
twins of the different pieces of equipment and aggregate them in Omniverse, that’s
something people are starting to recognize the value of because it’s a necessity.”
The criticality of domain expertise
All these advances are part of the evolution from automation toward autonomy, Lebaredian
explained. “Today, most automation is a fixed set of systems coordinated explicitly, but
we’ll soon have autonomous components in these systems that can adapt to the world
around them.”
While the computing resources to do this exist, a critical component to make them viable is
the domain expertise that needs to be supplied to the AI system directing these
autonomous systems. Which is why the knowledge of the industrial workforce will remain
important.
“When people ask me what someone should study in college,” said Lebaredian, “computer
science is still important to understand how computing systems work, but what’s more
valuable now is having domain expertise. So, I advise studying physics, materials science,
pharmaceuticals, medicine. We’ll still need computer scientists, but the value of domain
knowledge is becoming greater than the knowledge of how to program a computing
system, because that will be taken care of with AI.”
Given these advances, Lebaredian said manufacturers should step back and look at their
company and products from the viewpoint that anyone in the company can be a
programmer.
“That’s why it’s important to build the right team with domain expertise,” added Moret. “The
pieces don’t just fit together on their own. You need the right team internally and externally”
to guide that.
