Brent Barcey, VP of Sales & Business Development at Plus One Robotics, joins the Manufacturing Matters podcast to discuss the latest innovations in robotic handling applications on the factory floor. Plus One Robotics takes a novel approach to random mixed depalletization and high-speed parcel induction and makes ease-of-use a top priority. This includes properly building the model upfront to accommodate a variety of label conditions.  Another example is Yonder, Plus One Robotics’ “human in the loop” software suite that trains its crew chiefs as remote supervisors who monitor and direct robotic work from afar to keep facilities operational 24/7. Barcey also discusses robotics in logistics and the idea of people acclimating to the idea of robots handling packages.

Winn Hardin: [00:00:00] Hey everybody and welcome to Manufacturing Matters. We’re in Boston today at the Autonomous Mobile Robotics and Logistics Show, which is co-located with the Vision Show. And I’m lucky enough to be here with Brent Barcey with Plus One Robotics, vice president of sales and business development.

Brent Barcey: [00:00:13] Yep.

Winn Hardin: [00:00:14] So Brent’s going to give us a little bit of his time to tell us about Plus One Robotics. So let’s start there. You guys have a unique business model, and I can’t say how excited I am about it. So tell us a little bit about it.

Brent Barcey: [00:00:23] Yeah. So I mean, we’re focused on two or three different really whiz-bang applications right now, but we’re an AI vision-based company. So people that come to us have problems with grasping a box, moving it from point A to point B. They’re trying to simulate all these type of things that happen in a logistics atmosphere. So we focus really on three applications at this point. We’re doing random mixed depalletizations. So everything that comes into the building comes in at a pallet and it’s got to get depalletized and inducted.

Winn Hardin: [00:00:52] For sortation and distribution.

Brent Barcey: [00:00:52]. Yep. Handling boxes very quickly. Packages very quickly. We’re very public with FedEx. There’s videos everywhere out there, but we do that best in class. And then third, which we’re showcasing here at the show, is robots or robot interaction. We’re loading and unloading AMRs. AMRs have a lot of sex appeal right now in this industry, and we’re coupling those technologies to be able to give a solution to the marketplace.

Winn Hardin: [00:01:14] So another part of what makes Plus One Robotics unique in my mind is this service component, this remote management monitoring, and I want to call it troubleshooting. It’s more like operational efficiencies. So can you tell us a bit more about that?

Brent Barcey: [00:01:28] I called POPD. It’s cheap insurance. We’ve got a human in the loop, right? It’s called Yonder. That’s our software. And a robot is never going to be 100%. There’s going to be times when the packaging set just changes. It’s Christmas, it’s Hanukkah, it’s Fourth of July, whatever it is. So that gives trouble for the vision system. It gets stuck because it doesn’t know what it’s looking at. We can phone a friend. We have a human in the loop called Yonder. So the robot says, “Oh my gosh. I’m in trouble.” It goes up to the cloud. Comes back down to a crew chief that sits in San Antonio. They monitor 40 to 50 robots at a time, and the crew chief can see what the robot sees. So if everything’s okay and for whatever reason the vision system’s had a problem, it can click that package and it can go back to work. That usually takes about four and one half seconds. So if you’re walking the floor, you would never know that a human intervened remotely. So it’s a really neat technology, and as it does that, the model’s getting smarter and better. So after the next couple of times that we make that pick, the model will know that box is good to go and it will pick that automatically. So we don’t need to continuously do that. So that’s why it’s probably less than half a percent of our picks where humans intervene. But when you need it, you need it, because you don’t want the system going down because you didn’t induct this into the . . . A lot of times that robot sits idle for 20 minutes and no one knew because they’re busy in the facility. So it’s pretty, pretty awesome technology.

Winn Hardin: [00:02:47] So normally when AI or deep learning companies talk about human in the loop, it’s going to be during the tagging, the training of the initial model. You guys are bringing it in the operational mode, but you’re also continuing to always optimize and improve your model, and you’re working in the cloud environment. Most people think, I’ve got to come up with a really solid dataset. I’m going to create the best model I can. I’m either going to be able to do anomaly detection from that or I’m going to be able to do some very good inspection or some guidance based on this existing model. But what people have a tough time with is to adapt to additional product variations and expanding that model, especially on the fly. So how does Yonder deal with that?

Brent Barcey: [00:03:26] It’s the exceptions are the rule, and that’s where most companies, they get that model to a point where it’s perfect and then someone changes, and everything just goes to hell, right? Where we kind of pride ourselves in the exception handling. That’s where this is going to be able to help you get the ball over the goal line. But like you said, in that scenario you just mentioned, you have to train that model for however long in order to be able to update it. And then you can go. With Yonder, the end user does not have to give us the packaging of the SKUs. We continue to run. It doesn’t stop. And I think that’s where the differentiator is that we pride ourselves in is that we continue to keep your facility up and running. If we have to use Yonder to do that, that’s great. I’d rather do that than have to pull people off and do model updates, and it’s going to be clunky. I’d rather be clunky on the fly than clunky and just shut the system down. We don’t want to do that. We want to keep your system running.

Winn Hardin: [00:04:23] Your training procedure that you go through when you’re training the initial model, how does that work? I one of the difficult times in standard manufacturing of course is how do I get cameras? How do I acquire enough images of defects or common goods or whatever to build a semi-bulletproof model?

Brent Barcey: [00:04:36] Yeah, that’s a good question. Most people when they send you the SKUs, they’re usually perfect. Well, that’s not real life. Really pallets come in. They’re banged up and all that kind of stuff. So usually what — not usually — our process is: you’re going to send us the SKUs, we’re going to build that model, we’re going to take different images of it. We’re going to bang the boxes up. We’re going to make sure that the model understands that if this pallet of vodka comes in and it’s smashed in on one side, if you can still pick that safely, you should be able to do that. So we spend a lot of time up front with that model. It doesn’t take very long to get it up and running. But once we do the Yonder and the tagging and all that kind of stuff, auto updates, it goes fairly seamlessly and quickly.

Winn Hardin: [00:05:21] Out of all the folks that I’m talking to in deep learning and AI, which is a lot of them, you hear simulated effects or simulated conditions, working with initially, where we find one defect, we move it around the box or something. And that’s going to make our models stronger, which it does. But I’ve never heard one say that they actually physically just get up there and beat the crap out of the thing. I think that’s smart because we should all be thinking about luggage handlers, right? I mean, if we’re trying to build a good model of ATR systems in an airport, then we need beat-up pieces of luggage, with duct tape on it and pearls wrapped around the thing.

Brent Barcey: [00:05:56] Not everything’s going to be a brand new to me. Yes. Yeah.

Winn Hardin: [00:05:59] Yeah. So tell me a little bit about this Bad to the Bone demo. We’re seeing it on the show floor. You’re bringing together a lot of different technologies.

Brent Barcey: [00:06:05] Yeah, so we’ve got an arm. We’ve got our vision running it. And then we have a fetch AMR that’s running around with a couple of totes. There’s a donor tote and receiver tote, and it’s pretty neat. We’re picking out of the one tote and ducking into the next, and it’s kind of driving around the general population that’s in here and it’s hooking back up to the conveyance system and bringing it through. It’s just to kind of a showcase whether it’s parcels. I think most of it’s all clothing in there at this point, bagged clothing, but just showcasing what this could look like in in your environment.

Winn Hardin: [00:06:37] So you’re not even having to operate just in the demo area either. You’re in the aisles and operating it, right?

Brent Barcey: [00:06:42] Yeah. Yeah, we are. And to kind of mess with the system a little bit, the engineers will throw some stuff in there to trick it so we can show people, it’s running so well that it’s not going to Yonder, which those are champagne problems, we understand that, but we’re throwing different things in there just to kind of trick the system so people can actually — and it’s great here too because it goes up to the cloud. I mean, we’re connected to the cloud, so when somebody runs it remotely, they’re running it and they’re running the robots. That’s pretty neat.

Winn Hardin: [00:07:09] Just out of curiosity, we talk a lot about 5G and the impact on our space and industrial automation. And generally speaking, most of the people that I speak to are still like, you know 5G is a wonderful thing, but it’s intermittent. It’s not bulletproof. But I’m wondering in your space, is that something that could actually really be a benefit?

Brent Barcey: [00:07:27] It will, but not yet. I don’t think we’re there yet. We still like the hard wires and like the cloud connection the way we like it. I think 5G is going to get to that point. The customers that we deal with are really ready to rely on it as a sole source.

Winn Hardin: [00:07:44] I can see that especially in industrial or large environments, which are probably the type of customer that you’re targeting primarily. You mentioned FedEx earlier, obviously one of the largest carriers on the Planet Earth. But I wonder if 5G will be there when you have smaller facilities, smaller retail, back office, you know, RMA-type systems.

Brent Barcey: [00:08:03] I think it will. And I think in the bigger facilities and the more populated areas, we’re still blazing trails. I mean, if you really look at robotics as a whole within the logistics arena, there isn’t a whole lot of robots that’ve been deployed. We’re still blazing this trail. So I think it’s going to take a couple of years for people to get acclimated to the idea of robots handling packages in the facilities. And then once you start getting to these remote areas, I think the comfort level will be such that 5G will make sense for a lot of these people.

Winn Hardin: [00:08:32] Are there any discussions about actually marrying some of these robotic arms directly on the AMRs, AGVs? And is there a position for Yonder in that?

Brent Barcey: [00:08:39] I think there’s a couple of people looking at doing that. A lot of times you have to remember just because it’s a collaborative robot doesn’t make it a collaborative application, right? We have still yet to see someone successfully come up with a TAM that makes sense to be able to affix an arm to an AMR that’s going out into the facility. Really, for a human to have this tranche of space to be able to say, I’m going to pick your pick here and put it on the LocusBot, hit “go home,” and it goes to the pack-out area, that’s still the most efficient way to do it. With humans.

Winn Hardin: [00:09:16] So before I asked you about the Boston show, what you’ve seen.

Brent Barcey: [00:09:20] Yeah.

Winn Hardin: [00:09:21] I’m just fascinated with your product and your go-to market strategy and the value props you guys are bringing. What’s the next big thing that we should be looking for for Plus One?

Brent Barcey: [00:09:31] Our roadmap is really strong right now, and we’re going to have a product release next year that’s going to be a big deal. I can’t really get into it, but it’s going to be a big deal. It’s going to be a very public big deal.

Winn Hardin: [00:09:43] And we’ll just be on the software side? Hardware?

Brent Barcey: [00:09:46] A marriage of both, right? You know, at the end of the day we’re a software company. We’ve been forced to deal in the hardware business. So we play with a lot of hardware. The end users want us to do the first article and the pilots. Where the risk is part of the whole sell, right? The vision. So they’re like, no, Plus One, you’re the ones that are going to integrate this install, and then the integrators take it over from there. So yeah, we’ll be the ones marrying a couple of technologies. It’s going to be an awesome product. I’m really looking forward to it. Probably end of Q2 next year.

Winn Hardin: [00:10:14] That’s fantastic. How big is your company right now?

Brent Barcey: [00:10:17] We’ll click 100 people I think in the next two weeks, so we’ll be at triple digits. Yeah.

Winn Hardin: [00:10:21] So you’re obviously having to do a lot of consultative sales, especially at the OEM level, but are you doing a lot of handholding through the integration too?

Brent Barcey: [00:10:28] Very much so, yeah. Again, a lot of these people, this is the first time they’re deploying a robot in their facility. So we need buy-in from the maintenance team. We need buy-in from people on the ground. So yeah, I wouldn’t necessarily call it handholding. It’s just an educational process that they’re all going through together at once. It’s pretty neat.

Winn Hardin: [00:10:48] Fantastic. Cool. So last question.

Brent Barcey: [00:10:51] Yeah.

Winn Hardin: [00:10:51] Other than what you got going on a Plus One, what’s the coolest thing you’ve seen on the floor?

Brent Barcey: [00:10:56] You know . . .

Winn Hardin: [00:10:57] I know you’re near auto.

Brent Barcey: [00:10:59] The AMR technology to me is just awesome. I mean I co-oped at General Motors, and we had all sorts of vehicles running around the building. But what they’ve been able to do since the 25 years I co-oped at GM is just awesome. The autonomous forklifts, all the things that you’re seeing here. I mean, autonomous forklifts coming from General Motors. I mean, you wouldn’t have even dreamt that up.

Winn Hardin: [00:11:21] Yeah.

Brent Barcey: [00:11:22]. I mean, you’re watching forklifts load and unload stuff. Yeah, it’s awesome.

Winn Hardin: [00:11:28] All right. Thank you so much for coming and sharing your time with us today and telling us about Plus One robotics. I’ll look forward to a lot more conversations as we go forward, because I’m looking for a lot more innovations from you, and I’m gonna keep an eye out for next year. You’ve got to give us a little bit of preshow. We’ll be ready to figure out, at least understand what you guys are doing. So in the meantime, thanks for joining us on Manufacturing Matters. If you’ve got any questions for Brent or Plus One Robotics, shoot them on over to Tech B2B, share, like this video if you would. And we look forward to catching up with you again. Thank you.