Bonus Episode: Inside a Remarkable Factory

Host: Hello and welcome to this special ‘bonus’ episode for subscribers to “Technology Powers X.” It’s a story about how robots at the edge are helping manufacturers tackle labor shortages and keep their workers safe while scaling their business and outperforming the competition.”

On a country road, on the edge of town, in a rural part of the southern U.S., there’s a large, low-lying building. You don’t really see it from the road. There’s a row of trees blocking the view. And even if you did see it, you’d probably just keep driving.It’s just another industrial facility, an assembly line churning out widgets or whatevers. 

Except it’s not. Because what’s inside the factory is actually quite remarkable. What’s being made here are engine components for commercial jets. To be specific — they’re making turbine shrouds out of Ceramic Matrix Composites. These engine housings can handle temperatures of more than 24-hundred degrees, and they weigh less than a third of metal parts they’re replacing.

This single part is making flying 1.5% more fuel-efficient. Multiply that by all the planes in the sky and all the miles they fly, and you are talking about saving millions of dollars in fuel costs and millions of tons of carbon emissions. But as cool as that is… it’s not what they are building that I’m most interested in. It’s how. 

The “how” is by using robots. Not just fixed assembly line mechanical arms that repeat the same motion over and over again and stop everything if they break down outside of maintenance cycles — that’s old-school. These robots plan their own tasks, adapt to changing situations, and interact naturally with their human colleagues. 

Before these new robots came to town, the plant would turn out 50 units a week.They are now on track to produce 15-hundred a week by early next year. They’ve retrained staff to make that happen too. People now spend their time doing higher value tasks, while the robots do the heavy lifting. We call this factory “remarkable.”

Now we can’t describe in a single podcast all the robots at work, so we’re going to zero in on just one type. The ones we’re interested in, well…. They aren’t the two-legged chatter boxes we see in kid’s movies. And they’re not big-muscled chrome-plated menaces stomping down city streets set on eradicating the human race. Nothing like that. They are, well, let’s say, “understated.”

They are low, flat platforms that glide through factories and warehouses carrying stuff that would otherwise have to be lifted, or pulled, pushed, or tugged by humans or moved by a forklift from where they are to where they’re needed next. These smart platforms on wheels are called Autonomous Mobile Robots, and they are made by OTTO Motors.

Ryan Gariepy: An autonomous mobile robot is a vehicle which is capable of safely navigating real-world environments and all their unknowns.

Host: That’s Ryan Gariepy – the CTO of OTTO Motors.

Ryan Gariepy: So that’s people, that’s forklifts, tricycles, bicycles, everything.

Host: Their robots help manufacturers tackle labor shortages, scale their business, and outperform the competition. This is “Inside a Remarkable Factory” – a podcast about making manufacturing facilities:

• More efficient
• More productive
• And safer

This is the next great productivity revolution, and it’s brought to you by Dell Technologies and Intel. We’re going to find out from Ryan how his team is helping factories around the world achieve remarkable outcomes with artificial intelligence and automation at the edge, and from Dell Technologies and Intel about how this is part of a bigger trend toward smarter manufacturing with edge computing and 5G networks.

Step one is to figure out exactly what a modern factory needs. 

Ryan Gariepy: Generally, it’s about flexibility and efficiency. They’ve got workflows that they understand already, and now they’re able to repurpose their team members, their workers into other areas.

Host: And these robots get better by the day.

Ryan Gariepy: The previous generations of robots were limited. They didn’t have freedom. They were constrained to follow a strip of tape on the ground, or they were robot arms stuck within a cage. But now you have these robots which are able to re-plan and move around people and move around forklifts and make their own decisions in how this material moves around. And that’s a completely different dynamic.

Host: OTTO Motors’ robots use Artificial Intelligence (that’s AI) and Machine Learning (that’s ML) to adapt to their environment quietly, without fanfare.

Ryan Gariepy: That’s the sort of thing where AI and ML doesn’t always have to be this sexy, big picture thing. If the robot sees that a dock that it traditionally thinks is in one position has actually moved over a foot or two feet, because someone has done factory construction, the robot will figure that out.

This kind of flexible automation often yields big results. For example, a robot might learn to take a slightly longer route with less traffic so it can move faster. Sometimes it can just improve the robot’s performance, improve the robot’s average speed, 5%. And 5% makes a difference, right? Like if you went to any automotive company and said, we can run your factory lines 5% faster for no additional operational costs, they would love you for it.

Host: But it’s not just about speed. 

Ryan Gariepy: These robots have a growing capability to sense their surroundings, so now there’s new potential to optimize anything the robots can see. That can be safety related. That can be process control related. That can be inventory control related. There’s all sorts of potential here, because now you have these immensely powerful edge computing devices that are actually moving around.

Host: All that adaptability is showing up on the bottom line. That has changed the way manufacturers view AMRs.

Ryan Gariepy: The question we were asked five years ago was, okay, does it work? Is it safe? The questions we are asked now is how much does it cost for my application and how much money will it save me? These are very different questions.

Host: OTTO is a great example of how manufacturers are using data on the factory floor to improve their outcomes. In OTTO’s case, the analysis is done onboard, in the robot’s “brain,” which has Intel-based architecture inside.

But not all of the devices on the factory floor have their intelligence built in. That is where edge computing comes in. While there are many opinions on what the “edge” is, at its core it is where businesses collect data and use it immediately. To do that, you need some kind of edge computing–either onboard something like OTTO or infrastructure nearby capable of instantly collecting and analyzing data from devices, controllers or sensors already on your factory floor.

Todd Edmunds: Edge computing is being able to take those smart factory applications, smart factory platforms and the workloads that are generating efficiencies and being able to run them at the source where data is being generated, where that information is needed and where real time actions happen without having to wait for the latency of the cloud.

Host: That’s Todd Edmunds, the Global CTO of Industrial IOT and Edge for Dell Technologies.

Todd Edmunds: I think it’s becoming fairly apparent in the industry that ,whereas cloud was the initial easy button, you get more advantages because of that compute infrastructure at the edge where the data is, where the action is happening. You have a lot less latency. You have a lot less problems with data sovereignty, a lot less security problems and even a lot less cost by moving a lot of those workloads to the edge.

Host: So rather than collecting data, transmitting it, and storing it so you can analyze it somewhere else at a later date, you do the analysis in real time–right at the edge.

Todd Edmunds: And because of that, the data and how much data and what types of data that you use at the edge are starting to explode, too.

Host: That explosion of data is real. It leads to all sorts of issues. So, Todd has spent a lot of time figuring out where to put it. He has identified two ways that manufacturing customers are handling data at the edge.

Todd Edmunds: One is the typical data domain that everybody’s been used to for so long: spinning media or even solid state media hard drives that we all know and love in our, in our laptops, in our desktop computers. The second type that’s starting to emerge is a much more enterprise-grade hybrid type cloud data that stores and runs enterprise applications, but also allows that access to that data from anywhere.

Host: Analyzing data at the edge is efficient. But you probably will still want to store it in some way. Some of that data has even greater value when it is combined with data from sources outside the factory. That’s where the cloud storage or the hybrid cloud storage solutions Todd is talking about come into play. It’s a no-brainer that if you were building a factory from scratch you would incorporate intelligence at the edge every way you could. I mean, that’s the ideal scenario.

Todd Edmunds: So the key is to make sure the infrastructure and the architectures that are going to be inside the plant from day one, to be able to run all of the applications and use cases that a customer may have today, but also make sure that it’s a scalable enterprise grade infrastructure to be able to run the applications that we haven’t even thought about, in the future.

Host: The best way to “future-proof” your infrastructure is the use of a 5G network. That’s a cue for Caroline Chan, Vice President and General Manager of Network Business Incubation at Intel. 

Caroline Chan: 5G is the current wireless standard, but unlike the previous generation, LTE and 3G, 5G is designed for the ground up beyond the consumer space. It also enables enterprises to conduct what we call digital transformation, such as bringing AR and VR to your factory floors–all the things we wanted to do before, but now we are able to do it wirelessly. 

Host: 5G brings increased bandwidth, but its most notable quality is the decrease in latency. That is enormously important in a manufacturing setting.

Caroline Chan: For example, if you have a robotic arm that you want to very precisely cut through sheet metal to do a very precise operation it requires a low latency and high throughput type of communication link. That’s what 5G brings to you. It brings your capacity. It brings your low latency. It brings you a very deterministic results.

Host: Earlier, Ryan from OTTO Motors talked about being adaptable: That’s the idea that factories need to be able to change what they do and how they do it. A 5G environment brings that flexibility.

Caroline Chan: So, for example, in manufacturing today without 5G, they would likely be relying on some ethernet cables, but that means that you have to refactor your factory floor every time you’re changing your manufacturing production line. But with 5G, you’re able to do this wirelessly. It significantly reduces the cost and the time, and also gives you a much finer knob that you can turn to adjust the capacity and the throughput to exactly what you wanted.

Host: But security, privacy, and data sovereignty concerns also come into play. And they all add up to the need for complete control over where your information is, and who has access to it.

Caroline Chan: Many times, when you do things like machine learning, you have a tremendous amount of data. You don’t want to send all of that to the cloud.

Caroline Chan: So, typically it’s both a control and security issue that really meet your own business requirements.

Host: Speed, flexibility, and privacy are great, but the most important thing about 5G is that it allows you to dream big.

Caroline Chan: 5G is a tool that can allow you to significantly make your factory floor remarkable far more than just an incremental improvement. It becomes a transformational effort. I think enterprises can implement 5G in many ways. Number one thing is to select the right partner. 

Host: That might mean outsourcing the whole process to a telecommunications company. Or you can partner directly with Dell Technologies, which incorporates Intel technology in infrastructure that enables 5G networks.

Caroline Chan: That’s the nice thing about 5G. It gives you the flexibility to adjust based on your budget, your requirements. And we see a lot of these factory floors that are either DIY or hybrid, or completely outsourced that to the operator.

Host: One of Ryan’s customers sent him a video of OTTO’s robots in action. And the robots were doing robot things. But what really made him smile is what he saw in the background.

Ryan Gariepy: It’s a facility where there’s about 20 of our AMRs and they’re just moving boxes around. But the thing which is really compelling is that you don’t see people rushing around and pushing boxes. You see some maintenance staff walking around, but it’s more casual that you see them walk up to each other, you see them talk to each other, you see them go back to their tasks. There’s not people running freakishly up and down trying to get things done. It seems a lot more pleasant.

Host: Incorporating robots into your process isn’t a plug-and-play solution. It takes strategy. That’s because the benefits of smart manufacturing don’t come from making individual machines smarter, but from gaining continuous insights into, and control over, every part of the process, from the start of the supply chain to the end customer.

Ryan Gariepy: If all you do is buy some robots and throw them in your factory, there’s a risk of it becoming a science fair project. You’ve bought a handful of robots, and that was your thing, but they didn’t actually provide any business goals because your team didn’t know how to work with the robots.

Host: That’s why OTTO Motors does more than just sell robots.

Ryan Gariepy: So, what we also do aside from the robots is we provide fleet management systems. We provide a data pipeline and we also provide applications, engineering and simulations and, you know, digital twin functionality to ensure that these enterprises are getting more than a science fair project.

Host: One of the most promising things on the near horizon is that OTTO’s robots will soon be able to manage their own maintenance, making the factories they work in even more efficient.

Changing a factory from simply modern to “remarkable” is not a simple process, but it will probably include simplifying your edge and using 5G in ways you hadn’t previously considered.

Host: Here’s Caroline Chan about the evolution of 5G.

Caroline Chan: Very early on we were looking at 5G as a digital transformation, more than just a communication tool. So, we set out to enable an ecosystem that is cloud native, that’s virtualized, and because the Intel business has always been from the cloud into the network and to the end devices like the IOT devices. So we understand the entire ecosystem. You want to make it as consumable and easy to deploy as possible. So, we really look at ourselves as the ecosystem enabler, the Intel inside principle, extended beyond what you typically see in the laptop, but into the 5G network. So that’s how we see our role.

Host: And Todd Edmonds on how to simplify your edge.

Todd Edmunds: So Dell can help with three main things. Number one, we have a phenomenal team of people across the globe who know this scalable enterprise grade infrastructure approach through and through. Number two, is to make sure that you build security in from the beginning: embedded from the ground up. No longer is security just something you can bolt on. It has to be embedded into the infrastructure and into the solutions from day one. Lastly, use your Dell representative to help you plan that scalability and manageability piece so that you can repeat that across the factory or across all the factories across the globe.

Host: It all adds up to improved efficiency, productivity, and safety. And those three things all contribute to the bottom line. I can’t help but wonder as I watch a video of OTTO cruising through a factory, its little turn signal lights blinking away, communicating in its own intuitive way, whether people become attached to their OTTOs.

Ryan Gariepy: We are really interested in the field of human-robot interaction in general. And there’s just some interesting things, like why do people name our robots, but not name their own forklifts? But most of our customers named the robots, we do not ask them to name the robots. We don’t tell them that other people named the robots, but everyone ends up naming robots.

Host: Now, I don’t know if they’ve named the robots at that aviation factory I talked about off the top, but just in case they haven’t yet, let me suggest: Charles, Amelia, and Chelsey–as in Lindbergh, Earhart, and Sullenberger.

To learn more about this episode, visit DellTechnologies.com/TechnologyPowersX.