For years there’s been talk about solving the “last mile” challenges of getting goods and people to their destinations. Now University of Washington researchers are zooming in on what they’ve dubbed the “final 50 feet.”
A UW team is tackling the last link in the supply chain for urban deliveries, which includes searching for parking, moving items from the truck and navigating a route across traffic, sidewalks, bike lanes and building security to the recipient. A report from the group, which is part of the UW’s Supply Chain Transportation and Logistics Center, estimates that 25-to-50 percent of the transportation supply chain costs are driven by that last phase of the delivery.
And the situation is likely to worsen.
UPS, FedEx and Amazon combined are shipping roughly 10.2 billion packages annually in the U.S., according to a study released last month from the investment bank Morgan Stanley, as reported by Forbes. But that number is expected to spike to 14.9 billion packages by 2022.
Amazon is pushing the system harder than ever with its commitment to one-day-delivery made in April 2019. The quick turn on orders is driving up costs for Amazon and cutting into profits. GeekWire estimated that shipping costs could total more than $35 billion for the company last year. And as Amazon promises more, other retailers work to match them.
Meanwhile, cities are becoming more densely populated with people, vehicles and bikes, while there’s mounting climate-driven pressure for environmentally friendly solutions.
So the center has created a public-private partnership called the Urban Freight Lab, which is studying the challenges and testing solutions to make the last stretch faster and more energy efficient. The researchers are collecting on-the-ground observations, GPS data, and information gathered from sensors that will be placed in commercial vehicle loading zones to better understand the situation.
Their proposed improvements include an app to help drivers find empty loading zones and speeding up drop offs by installing shared delivery lockers in easily accessed, public locations such as alleys, parking lots and curbside locations.
They’re novel fixes to a problem that’s gotten limited attention, the researchers said.
“It’s a very costly system. It’s expensive to do the last mile, but no one in the existing framework can take it on and fix it,” said Anne Goodchild, lead researcher on the project and founding director of the UW’s Supply Chain Transportation and Logistics Center. One of the big hurdles is that so many different parties are involved, namely cities, delivery companies, consumers, residents and property owners.
The three-year project has $2.1 million in funding, including a $1.5 million grant from the U.S. Department of Energy. Other partners include the Pacific Northwest National Laboratories, which is working on the app for drivers, and Seattle Department of Transportation. The effort just finished its first year. The UW lab has also investigated the impacts of ride-hailing vehicles such as those run by Uber and Lyft.
One of the group’s first steps was shadowing drivers with UPS and Seattle-based Charlie’s Produce. The researchers were wowed by the intensity and scale of the work. A driver might start the day with a truck packed with 300 parcels and 100 addresses to hit. The ride-alongs revealed the strategies that drivers are currently using, and their real-world challenges including people trying to nab food from the back of the produce truck and commercial zones filled with passenger vehicles.
“We rode with very experienced drivers so we really learned how amazing it is, how they manage,” said Giacomo Dalla Chiara, a UW research associate.
“They don’t stop for anything if they can help it,” said Lyndsey Franklin, a user experience research scientist with PNNL. “I was surprised they put the truck in park half the time.”
A UW analysis of GPS data shared by commercial delivery companies showed drivers are spending 2.5 minutes on average finding parking. When repeated over 20-to-30 trips in a day, that’s an hour of wasted time and gas spent looking for a spot to pull over.
As part of a pilot project, the partners are focusing on an eight-block area of Seattle’s Belltown neighborhood. Over the next year, they’ll be installing sensors at loading zones covering roughly a mile in length. Because of privacy rules in Seattle, the sensors can’t be video cameras, but will instead be magnetic and light sensors to detect vehicles. They’re also placing lockers at multiple locations to save carriers trips into buildings.
PNNL has already completed a first version prototype of the app. The plan is to create a final app that can tell in real time if there are open spots in load zones based on sensor data, as well as using machine learning to analyze past data to predict the likelihood of a spot being open currently or by the time a driver will arrive to a destination.
The longer-term objective for the overall project is to figure out which elements would be scalable to a citywide application, as well as for other municipalities and delivery fleets. The researchers shared their eagerness to make a positive difference.
“We’re not just building this thing that is going to dump out the output,” Franklin said. “We’re helping that guy from the produce company avoid circling the block. Or the poor delivery guy who has 10-to-12 hours of work in the back of this truck not have to sprint so far. There is a very human piece. These drivers are really working hard.”
Editor’s Note: Funding for GeekWire’s Impact Series is provided by the Singh Family Foundation in support of public service journalism. GeekWire editors and reporters operate independently and maintain full editorial control over the content.
