Researchers at the Department of Energy’s Pacific Northwest National Laboratory are contributing their expertise in artificial intelligence, machine learning, and app development to a project that will ease challenges with urban freight delivery, an experience especially difficult during the holidays.
The entire project—funded with $1.5 million by DOE’s Office of Energy Efficiency and Renewable Energy’s Vehicle Technologies Office and led by the University of Washington’s Urban Freight Lab—will develop, test, and improve technologies aimed at cutting time spent by the driver at the curb, increasing productivity, and reducing time and fuel spent searching for available parking.
EPISODE NOTES
Intro:
Welcome. I’m your host, Jess Wisse. On today’s episode we’ll talking about how something that may surprise you. Here are a few hints: the holidays, online shopping, and parking.
Stay tuned to learn more.
JW: If you’re anything like me, I’m sure you’ve been busy preparing for the holidays. This includes wrapping gifts, RSVPing to parties, and online shopping. LOTS of online shopping. But have you ever considered all of the hands who’ve touched your latest shipment before it arrives on your doorstep?
PNNL user experience scientist, Lyndsey Franklin thinks about this daily.
LF: Maybe I should set out, you know, Gatorade and some snack bars. No, forget giving milk and cookies to Santa. Give Gatorade and snack bars to your poor delivery drivers because they are hustling. It’s insane!
JW: Researchers at Pacific Northwest National Laboratory are using their expertise in artificial intelligence, machine learning, and app development to ease challenges with urban freight delivery, an experience especially difficult during the holidays.
Meet one of the researchers at PNNL working on this.
LF: I am a User Experience Research Scientist in the Visual Analytics Group at PNNL. I try to make computers better playmates for people. I kind of take the philosophy that if something goes wrong, it’s the computers fault. It wasn’t designed well. Or people didn’t think about it well, and it’s really not the person’s fault. Somebody needs to make the machine play better.
JW: Lyndsey is working on a project that’s funded by DOE’s Office of Energy Efficiency and Renewable Energy’s Vehicle Technologies Office. The project is led by the University of Washington’s Urban Freight Lab.
Lyndsey’s goal: to develop, test, and improve technologies aimed at cutting time spent by delivery drivers at the curb. Think of your average UPS or FedEx delivery driver. Lyndsey’s job is to increase their productivity and reduce the time and fuel spent searching for available parking.
LF: What if you could make something like parking in downtown Seattle smarter? The particular problem that they were trying to address was, ‘How do we make that delivery process more fuel efficient in crazy environments like Seattle?’ So, they came to us in two capacities. They were looking for some expertise in the modeling aspect. John Feo here at PNNL is leading up that part of the of the effort.
The other thing that was important to them, and very insightful of them, was to realize that this wasn’t going to be something they were designing for a typical desktop environment. This this wasn’t build a model, have it run on a big heavy machine, and spit out an answer, and give it to drivers. Downtown Seattle and the parking in downtown Seattle is ever-changing and always chaotic. And so, there’s this noisy busy environment and you’re supposed to be giving information to drivers who aren’t sitting at a computer. So, what is that whole experience that the drivers are going to have. What’s that going to look like?
JW: Finding parking can be a major headache for freight delivery drivers. Especially in cities like Seattle. Restaurants need a constant cycle of fresh produce. Retail stores depend on delivered products to maintain a steady flow of sales. People living in a city’s apartment buildings expect their Amazon purchases delivered on time, without fail.
That’s why PNNL is working with the University of Washington to develop an app that helps drivers identify open parking closest to a delivery location. The Urban Freight Lab calls this sweet spot for a delivery the “final 50 feet”—where a delivery driver stops to deliver their freight.
LF: So, the focus of the app is trying to help increase awareness for when parking might be available. In the case of newer drivers who maybe are seasonal, they’ve been added to routes to deliver Christmas packages and things like that. They’re not as familiar with the area. They don’t really have that internal map in their head of: ‘Well, if I can’t park here, I can park here. Or I could park in the super-secret spot that, you know I discovered by accident one day.’ They just don’t have that list in their head. So hopefully what the app can do then is bring more of an awareness of well, this is what’s available to you.
JW: But Lyndsey isn’t only trying to improve Seattle’s parking situation. Her primary objective is to create a tool that would actually help delivery drivers. Sounds simple enough, right?
But to create an app that’s actually useful, an app developer must first understand the people the app will help.
LF: The part that I get most passionate about is, you know, being able to help people realize that: no, you should question that. If you have to jump through all these crazy hoops to get your job done, there’s somebody in that chain who built those tools that didn’t finish the job. There’s more to be done. You don’t have to accept that.
You know technology is this weird unfriendly thing that’s just kind of foisted upon you and like you’re told, ‘Here. This is the tool you get to use. Use it or, I don’t know, do it by hand on paper or something.’ So, I like seeing when people realize that, ‘Oh, hey, I am the master of this strange beast we call technology and it’s supposed to work for me!’ So, I’m the people pleaser by heart I think.
JW: She needed to walk a mile (or two) in their shoes, climb a few flights of stairs, and drive all around downtown Seattle for a day. How did she do it? She went undercover.
LF: I got to actually you know dress up like a UPS driver, hang out in the truck with them and that experience was just wild. So the user experience I have it is once you have the information, and assuming you have data and a model, how do you present that to a driver in a way that’s going to actually help them and, you know, not interfere with a job that’s already crazy busy?
JW: We asked Lyndsey if she was surprised by anything during her ride along. Her response?
LF: So much! So much surprised us! We did two sets of ride-alongs. One was with a produce delivery company and the other was with UPS. So everybody’s pretty familiar with what UPS does, but I think maybe they are not as familiar with the absolute speed at which these drivers operate. They are flying. I cannot begin to describe just how fast they are sprinting from building to building particularly in Seattle.
JW: And the most surprising part actually had very little to do with delivery routes.
LF: For me one, of the surprising things was just how much building navigation there was, and some of it was really non-intuitive. I mean he would he would look at a box, read a label, and be like, ‘Okay, well, this is supposed to go to this this business. And you know, the address is this particular suite.’ But just based on having done this route for, you know, seven/eight years think he had been delivering.
And so, one of the surprising things for me was, if you are a new driver in some of these environments you can’t actually trust the label because they’re technically correct, but they’re also very wrong. And so for a new driver some, of the difficulty comes in just knowing where am I actually supposed to hand this box off to? These buildings that they’re delivering to are just—they’re mazes! And that’s on top of the parking issue.
I mean there’s not a lot of parking, there’s construction happening everywhere, you have inconsiderate personal vehicles who will, you know, park because, ‘Oh they’re just going to be in there for five minutes. And, you know, why shouldn’t they get to use that parking space too?’ But it’s like no, no, you’ve got people who really need those commercial loading zones. Stay out of the commercial loading zones!
JW: Another thing that surprised Lyndsey? The people side of the equation. It can sometimes be easy to forget that an actual human is responsible for delivering your packages.
LF: The produce folks had their own set of quirks. That one was amazing! They started at like three four o’clock in the morning to deliver to all of these commercial kitchens. And the gentleman that I was riding with had been on that particular route long enough. He knew everyone’s name from, you know, the back door to the front door. We’d be delivering produce and he’d be asking about how somebody’s mother enjoyed her vacation to Puerto Rico and they’re having, you know, full friendly conversations because they just they see each other day in day out. There was an amazing amount of interpersonal skill required for these jobs that I’m sure most people probably don’t even think about. But they are expert problem solvers. They are expert navigators, and they are sprinting all day.
You know I run marathons. I consider myself to be in semi-decent shape, and I was wishing I had brought like water bottles and Gatorade to keep up with some of these folks because not only are they just kind of sprinting around up and down stairs (because elevators take too long), but they’re carrying lots of packages while they do it. So, it’s just—this it’s not a job for the faint of heart.
JW: After getting a glimpse into what these drivers are up against on a daily basis, Lyndsey was inspired.
Even more than before, she wanted to create an app that could truly be useful to delivery drivers. The invaluable insights she gained from her ride alongs opened her eyes to just how powerful this app could be.
LF: Inspiration is probably a good word for it – because once you’ve done that process of actually kind of embedding yourself with them (I mean, like I said for UPS we had to actually wear the uniform) it’s a very different job from the one that I have and there is no substitute for actually experiencing it like that like we did with the ride along. So you come away from that going, ‘oh, you know, every expectation I had about what this thing might look like what it might do—toss those aside.’
JW: It was only after her ride alongs that Lindsey could envision a practical design for the app. She quickly learned that drivers wouldn’t have time to enter information into the app, an assumption she held before her ride alongs. And so, she adapted.
LF: They’re not going to have time to stop and input, you know, what their next stop and their route is going to be. We had originally had some thoughts of well, maybe we’ll recommend some parking spaces based on, you know, what their next stop is going to be—give them like a top three. They don’t have time for that. Even something as simple as, you know, move to the next stop on my manifest because sometimes there is no parking available.
And so rather than sitting and waiting they’ll just skip and go to the next stop and come back to it later. And so, the inspiration that we kind of started tapping into is: this isn’t going to be like getting directions from Google Maps. It’s more video game-like, in that you you’ve got kind of this, you know, the whole world is at play at once and you’ve got a player who is trying to, you know, make as many stops as they can in a short amount of time as possible. So, how do you lay out that information so that they can start to optimize in their own heads? With this additional information, and see what’s where’s my biggest value going to be? Because they just they don’t have time to input. It’s much more real time. You know, they are really working that hard. So, think less Google Maps, more like a video game. But it’s the it’s the sort of inspiration that you only get from experiencing it.
JW: Lyndsey also quickly learned that these drivers aren’t just master navigators. They are also excellent negotiators and teammates. All for the sake of getting the right stuff to the right people at the right time.
LF: There are, you know, chefs in the Seattle area who are very particular about their produce. So now you got to find parking. Now you got to carry packages around. And now you may have a chef or two in the mix who is unhappy with the quality of their mushrooms.
And so the gentleman that I was riding with, like I said, there was a tremendous amount of interpersonal skill. He actually knew when certain stalls in Pike Place Market were open, when staff were likely to be there. But well before the market was open, he even had strategies for where he could go to find the same item and a high enough quality that he could come back to this kitchen so that they can get up and running at four o’clock in the morning so that they can start serving customers.
And, you know, I don’t want to call it like an unofficial barter system, but there was there was a tremendous amount of negotiation that’s just once they get out of there their vehicle. There’s actually a surprising amount of camaraderie between drivers even of other companies and so what you find is the drivers actually start cooperating with each other to help protect the parking Our UPS driver would pull out into traffic in a certain way that it would block traffic so that this FedEx truck could come in and take the parking space right after him.
JW: The app that will help reduce double parking, blocked traffic, and parking fines takes all of this information, and more, into account. But for now, Lyndsey and her team are taking the time to focus on the prototype’s backend to make it as fast as possible.
LF: Not only do we have to present a lot of information, but it has to be fast. It really has to be fast. And so, we’re taking the time our main developer Amelia is doing an amazing job of very carefully deciding what the best services are for uploading and shifting data around in this platform. People look at this this mobile app that just appears on their device and they think all the all of the hard work goes into what you see. And really what you see is a dumb pretty window on top of some really sophisticated algorithms and technology in the backend. So right now we’re kind of working on making sure that back-end is feature complete and fast. And It may not look like much at the moment. That’ll come later because the hard work is in just supporting that pretty dumb window in the front end.
JW: So, while the app is still currently in the prototype phase, the Urban Freight Lab & PNNL have big plans for its future. Including installing sensors in an eight-block study area in downtown Seattle.
These sensors will collect data about parking spots and occupancy. PNNL will process and analyze the data and then feed it to the algorithmic model.
Historical data like truck size, type of delivery, and how long a vehicle stays in a location, combined with real-time data from the sensors will allow scientists to ‘train’ the model. All of this information combined will allow the app to tell delivery drivers when there’s a high probability that a parking space will open up.
LF: One of the things that we would like to do in the future then, is start understanding when drivers park in a particular space how many times do they go in and out of their truck? Is this a stop where people are likely to stay a really long time? So, there are actually some more delivery-specific behaviors that we would like to maybe think about how do we how do we fit that in?
JW: The need for packages to arrive quickly and reliably is only increasing; especially when we’re getting more and more accustomed to ordering everything online.
That’s why PNNL is excited to do research like this. Research that will researchers that will increase efficiencies and reduce fuel consumption. Research that will make people’s lives less stressful. Research that will make people happier.
LF: Hopefully we can make it less stressful for them to be, you know, bringing us the things that we’ve all gotten so spoiled that I can, you know, order with one click on Amazon and it’s supposed to be here the next day. And if it’s not, that’s tremendously inconvenient to me but you know that’s actually somebody’s job who’s hustling, and if we can make it so that, you know, not such an onerous grind to them then everybody’s happy.
JW: And with that we end this episode by wishing you a Happy Holidays, happy online shopping, and please, please do not forget to thank your local delivery drivers this holiday season.
Music
JW: Thanks for listening to Pods of Science. Want to learn more? Follow us on social media at PNNLab. We’re on Twitter, Instagram, Facebook, and LinkedIn. You can also visit our website at pnnl.gov. Thanks for listening.