Dr. Thomas Evans is the Robotics Chief Technology Officer for Honeywell. In his post, he sees the multitude of driving forces that bring automation and robotics into today’s manufacturing environment. We spoke to him at length on the current situation and the future in this realm, and how Honeywell is mirroring client needs with their offerings—and how those needs are changing.
Fifth Wave Manufacturing: Our guest today is Dr. Thomas Evans, he is with Honeywell. He is the Robotic Chief Technology Officer. Welcome, very much, Thomas. Thank you for joining us.
Dr. Thomas Evans: Excellent. Thanks, Dave. It was a pleasure talking with you previously. Glad we can get together and have some more dialogue.
FWM: Absolutely. Let’s launch into our questions here. First automation is not only here to stay, it’s really a necessary thing in the manufacturing environment. Give us a milestone. Where are we in terms of automation right now?
Evans: That’s a good question. I think you hit it on the spot: Customers are becoming more savvy and more intelligent as they’ve progressed through their evaluations and their view of the return on investment for adding automation. Increased adoption has happened in the last couple of years and we are seeing increased performance in robotic pick and place mobility, robotics to move pallets, cases, or raw material from point A to point B and complete the applications inside of the warehouse and manufacturing facility. We’re seeing a broader level of maturity and an increase in operational time that’s allowing us to get broader insights of how we make that automation even better than what it is in the initial adoption phase.
That’s something I’m excited about. Our customers are coming to us and asking the questions around not just putting that individual piece of automation on their floor, or increasing some productivity or a manufacturing process, but talking about the vision side of it. They are talking down the road; if that is my first step, where am I going to be in my roadmap for steps 2, 3, 4, 5 within the next three to five years? And in turn, what does that do for me? I need to know in metrics of productivity, throughput, yield, or even reduced dependency on labor. So those are the conversations that I’m really engaged in with our partners and customers. That is the current state of maturity of advanced automation and robotics, and this is projected against our own roadmap and our development where we’re going to be in the next three, five, and 10 years.
FWM: When you’re out there talking with people, do you think people were prepared for the explosion of robotics? I mean, everyone was saying, we don’t have any people. We can’t recruit the people. Now suddenly here’s this thing that can solve a lot of those problems. Were people ready for that or maybe are they ready for that?
Evans: I think they were ready, but there wasn’t the deeper understanding of taking the readiness position to a deployment position. There’s technology. You can go to many of the conferences, trade shows and see some great technology. There’s some cutting edge companies out there and algorithms and solutions that we see that are appealing. And we also do that in our development team organically, but they were not ready in the aspect of, okay, I love that technology. I can get an ROI and a productivity from it, but then how do I make it real inside of my workspace? How do I train my operational staff that’s currently there? How do I make sure that the operations staff that’s feeding a robotics or automation system is prepared to meet those specifications? We must get that return on investment in the yield after using it for shifts and months and quarters. I think that was the process. And looking through not just the technology, but who can be the solution provider to not only deliver the technology—mature technology—but also handle it from end to end in the deployment process, adoption process. Then it must be scaled after the success of the first adoption.
FWM: I think it takes a special person to do this. It’s not like saying, “Hey, we need some resources, let’s go buy another dump truck.” We have to plan it out, we have to fan it out. I think this is its own art. Yet you see it every day, I guess, right?
Evans: It is its own art. And I think we gain a lot from being in lockstep with the customers that want to adopt our automation and solutions. I encourage that with our product engineering and R&D team, to get completely close with the operators, the engineering teams that have interest or purchase our products or see us as a solution provider. That helps us to be in lockstep with what success means for them. First, that helps us to look at the products and the specifications that we have, and then make sure that’s compliant in their facility with their product, with their operations, with the manufacturing process. And second, once we get to that point of performance and the customer sees that productivity, how do we maintain it? So they continue that for multiple years down the road, but then look at the next step for where they can gain a competitive edge in their market.
FWM: AI is also here. It’s coming out and Mitsubishi uses it in inside one of its laser cutters, and there are a lot of things happening like that. In terms of AI what are you seeing out there and, and what do you wish would happen as AI unfurls for everyone?
Evans: It’s a very popular topic and has been for the last couple years. When joining Honeywell three years ago, I knew that we should invest and prioritize AI based upon the competitive position, especially the growth trajectory that we could offer to a broader customer base. Honeywell is a very diverse company. Yes, we operate in a lot of businesses, but artificial intelligence is one that can be applied throughout all our business ranks and through what we see in our broad customer base. The conversation has changed. We are deep in the algorithms, the neural networks and in the approach to make AI reliable in convoluted and complex settings like manufacturing and warehousing.
We have tapped into generative AI to draw inferences from not our automation systems just performing from point A to point B in a defined scope or space, but what happens and how to respond when the dynamic changes, or how you want to optimize a performance application or mechanism, drawing that inference from the two closest applications that were handled previously with intelligence.
That’s a simplified solution. In deploying robotics and automation, generally we are trying to replicate what the human system can do with sensing, feeling, visualizing, and then the cognition to be able to perform an intelligent action and then repeat the process or enhance the process. Machine vision is big and complex. You must make sure of some things in the environment when you are visualizing something. If you’re going to have data streams to input into an AI algorithm, that first visualization of what you’re seeing, what is it, where is it in the three-dimensional space and all that in various lighting conditions and dynamic motion tracking is important and complex.
We had to really understand the overlap between artificial intelligence, computer vision, and making that something that is reliable and scalable and harness that. This concept is coming out in our products and has come out in the last couple years. Our smart flexible palletizer uses it, it uses computer or AI-based computer vision to detect the complexities in the 3D environment of various pallets, whether it’s structured unstructured. The same is true inside our trailer loaders and unloads. We made them more advanced with the tools that are coming out with AI. For the products that already use AI, they are enhanced in new releases.
FWM: With these solutions, your data needs go up tremendously. There is no way to avoid these increased data needs. Is Honeywell working on some of the computing solutions that are that are new to the warehouse or some of the manufacturing spaces?
Evans: Yes, I can speak to that in regards to how we see it on the automation robotics side. I think that we’re very well positioned. You hit on a very good point. The data needs are a complexity for us. I mean, data stream for advanced computer vision, AI based computer vision solutions. We want to see every image all the time.
We want to have that in a global cloud architecture, which Honeywell is well positioned to do with what we’re maturing right now with Honeywell Forge [more information at Software for Digital Transformation | Honeywell Forge]; we now have broader enterprise architectures with complete connectivity to all of our assets. We now have a view at the machine level and a view at the software level with data insights and intelligence to the labor workforce. All of that is wrapped into an enterprise architecture, which we’re creating the data streams that will capture all the images from our AI-based computer vision solutions. With that we can draw stronger correlations with how we’re using the computer vision solutions in one application to another, or one customer facility to their next. That’s something that’s challenging.
If you look at the repository capability, Honeywell’s position better than anyone, and you look at our play to draw more deep correlations and inferences from that, I see that being picked up by quantum computing, and and we’re positioned very well in quantum computing. I mean, how do you optimize a very large data set that’s handling global data information? That’s one way you’re going to have to be able to compute for very large optimization sets. It’s something to be considered as it comes to global data modeling and, and global AI solutions.
FWM: That’s a great point. And in fact I was thinking as you said that there are those big companies that will need to balance their workloads globally, now you also need to have some kind of secure internet connectivity. And interoperability, you know, every time we make an advance, it seems like it climbs up the OSI [the Open Systems Interconnect model, more at OSI model – Wikipedia] stack, right? We go from physical connectivity all the way up to applications that need to interoperate across countries—yet another level of complexity. Do you have any customers who are facing that right now?
Evans: If you’re speaking about product security, we are active in that right now as far as utilizing more than one system communication protocol and driving that through our product security before the products hit the customer floor. It is a requirement across automation, product development, product engineering, and commercialization. If Honeywell is a lead position partner for a customer solution provider that’s going to be able to deliver the interoperability from a software perspective and also provide the same level of software security/product security, that’s within our DNA of what we provide in the commercialization process of product engineering. And we, we know firsthand the security threats, the risks, and how important it is to all of our customers, large or small, to make sure we stay ahead of it.
FWM: That’s a really important issue to a lot of people. We’ll get back to robotics for a moment. Where are people first deploying robots in your experience?
Evans: The applications we are now deploying are in the area of redundant tasks, or the task that is not highly productive for a laborer to be involved in. We are deploying in areas where we can provide the gains to increase productivity, sustain productivity over the shift or quarter, whatever it may be for the customer. These deployments remove personnel from redundant activity, highly labor intensive activity in difficult environments. You see a lot of that at the dock door. You see a lot of that at the palletizing side. Hence, we talked about some of the products unloading boxes inside of a container or a trailer in, let’s say, Arizona, with 120-degree heat in the middle of the day.
You see a lot of turnover in that area of the manufacturing facility, shipping and receiving. We have launched several products that are in robotic loading and unloading. We have those in commercial deployment in the last couple years and are maintaining and exceeding rates that the workforce was providing before we introduced automation. It is a common thread no matter what the vertical or industry. It’s the same situation for palletizing and de-palletizing, it involves redundant movements of 80- to 100-pound products. It’s a long shift.
We put a lot of technology and solutioning into the smart, flexible palletizer and some of our palletizing applications. Then again, we develop solutions in the articulated space, which is defined as robotic loading and unloading. If we’re picking and placing, then it’s the same with palletizing and then also mobility. You will not get productivity out of a worker carrying a product from point A to point B or from process machine one to process machine two. We want to make sure that can be handled in mobility and delivered.
Machine tending is a great example of that. We have deployed and delivered it in multiple facets, whether it be carting totes of product, individual items, or all the way up to bulk pallets or bundles of material. It is an area of adoption that our customers are keen on. I would say those applications I mentioned would be probably top two or three areas of deployment.
FWM: If with your permission, I’m going to tear up the rest of my questions because you’re a good speaker. Give us an idea, if you would, of Honeywell’s own roadmap. How do you look at the world? How is your deployment and customer need meeting? How is tht rolling out in the next, say, three years?
Evans: We look at the world pretty much through the eyes of our customers and deeply understand what they’re looking for to make their operations and business successful. That’s the bottom line. We dive into the deeper aspect of that product engineering or chief technology officer driving not just a solution and a product, but a solution path in lockstep with customers is the top of my responsibility list. So that’s how we see it as far as, you know, where is our roadmap in technology capability as far as what’s going to make a difference for our customer base. Then also we must be very adept and knowledgeable about the technologies in our IP portfolio that we’ve developed, maintained, and deliver new patents every day.
But also keeping an eye on the markets and new technologies that are coming in through innovations, through research and science that are going to impact where we’re going over the multiple years, five or 10 years. That’s how we create broader vision roadmap. If I’m speaking on robotics and automation on the robotics side, I look at what’s being done from various level technology providers. I am very keen on my team working in a sandbox in the R&D environment. If we want to push the envelope, we want to assess that not just theoretically, but in a test build environment to make sure we’re delivering something, and understand completely how it’s going to work in our vision roadmap over the next multiple years.
Once we get that ideation and that innovation for what’s next, Honeywell has a very rigorous product development path that we use. We have our executive and scientists coming in from other areas of our business, vet the concept to make sure it meets the requirements and is sound. And then it has to meet all those metrics and have the Honeywell signature on it before it goes in as a piece of the roadmap into production. So that’s the broad-level view and it’s a very challenging environment—one I enjoy working in. I think this is what differentiates Honeywell in our offerings.
FWM: I can imagine where a lot of the work is very solution-oriented work. So in running up against a challenge, maybe nobody’s invented this and we need a solution for it. Picking up soft things with a robot is one of those things. And several solutions are available. You probably have worked on that yourself, or one of your team members has. I’m trying to explore more because it sounded like you hinted at working ahead of the curve a little bit.
Evans: I think as a leader, that’s something that definitely happens as we’re driving in our direction to win and be successful and make our customers win. It’s going to push the envelope and it’s going to open up to the public and our competitors, you know, where Honeywell is going, that’s something that we want to just stay focused on. The teams across Honeywell are very much dedicated to it. But it also goes both ways. We see the changes in not just technology, but also changes in society, changes in areas with pandemics in which we have to be flexible and nimble. If we had a vision and roadmap, it, it can’t be fully set if everything around us is changing and changing quickly.
So that’s something I’m very proud of, is seeing how we have overcome some of those obstacles and, and kept an open eye to how those impacts affected our customers in the last couple of years. We really accelerated what we could do because we heard the demand for automation. And my team, and the other teams across Honeywell, we’re in the lab and on the customer site deploying and developing and maintaining at an accelerated pace. That’s what really captures the essence of being a global leader. Those are the things that really capture the public’s eye and the intention of what we do as a global engineering company.
FWM: What do you think are a couple of trends to watch for next year?
Evans: I think you’ll see broader adoption from some various levels in industry for automation. The general automation that may not be one that’s a cutting edge, but has been around for a couple years, and you’ll see more adoption rates. I think you’ll see more advancements in autonomous control, whether that autonomous system is a driverless vehicle, whether it’s a robotic pick and place solution or robotics mobility or the broader integrated solutions of those, the technology and proving out of that technology has matured drastically with the tools that are in our hands now. Third, I would say that not just that integrated solution, but also the use of a broad stream of data, and we talked about the data being so important to automation and the processes.
This broad stream across all of Honeywell also applies to our clients and we are capturing data, standardizing it, understanding how it can be reused. That’s what I’m very interested in, to take it from the point solution, integrated solution to more of a fully automated facility. If we get close to that percentage of a fully automated facility and just have the required amount of manual workforce to keep it churning, that’s a big win. But the way we get there is by truly understanding those foundational technologies and driving autonomous systems with a data stream that’s reliable, and then building on the capability of that autonomous system into full facilities. That’s what I see as the focus and some increased focus coming in the next couple years.
FWM: Well, great. Thomas, thank you so much. I know you’ve got a hard stop here but I really want to thank you for joining us on on our program and if you’ll entertain the thought, we’d like to have you back again sometime.
Evans: Excellent. Sounds good. Thank you.