Autonomous Package Deliveries May Not Reduce Emissions

(Inside Science) -- In 2018, nearly a quarter of surveyed Americans said they shopped online at least once a month, a fraction that has likely increased since the start of the pandemic. But as the demand for online goods has grown, so, too, has the need to deliver all these purchased items. Making these deliveries environmentally friendly, cost-effective, and accessible presents a big challenge.

By 2030, delivery vehicles are expected to increase by 36% over 2019 numbers across the world’s 100 most populated cities, resulting in a corresponding rise in greenhouse gas emissions. Increases in traffic congestion and decreases in parking availability could also contribute to the amount of pollution each vehicle produces. In addition, when there’s less parking, people tend to park illegally, idle in disabled spaces or otherwise disrupt access to sidewalks and public spaces.

To solve these problems -- and reduce labor costs -- companies are experimenting with autonomous delivery vehicles. Previous research on self-driving passenger cars has shown that because this technology is able to drive much more efficiently than humans, it could reduce traffic jams and, in turn, decrease pollution.

Automated driving can also improve fuel efficiency, said Gregory Keoleian, director of the University of Michigan Center for Sustainability. “But then on the downside ... you’re adding weight to the vehicle for the [electronic] equipment,” which could potentially offset any greenhouse gas emission gains, he added.

Plus, without a delivery driver, companies will need an alternative way to get packages from the vehicle to customers’ doorsteps. However, the production and operation of robots capable of doing this could further contribute to the system’s overall environmental impact. It could also cause congestion to move onto sidewalks .

If that’s the case, “that’s not a good strategy [to use] for the future,” explained Luyao Li, a sustainability researcher at the center. This dilemma motivated Li and Keoleian to examine the impact of these potential autonomous delivery services, starting with their carbon footprint.

In their analysis, which was published earlier this year in the journal Environmental Science & Technology, they calculated the amount of greenhouse gases emitted by several suburban delivery approaches. Their mathematical models compared a fully autonomous scenario, a fully human-run option, and a combined human-driven and robot-delivered approach. For all these scenarios, the researchers also looked at how emissions differed between vehicles with internal combustion and electrified engines, as well as between vans of different sizes.

They concluded that by itself, automation didn’t reduce carbon emissions. In fact, the fully autonomous method consistently produced either the same or slightly more greenhouse gases than when the packages were transported or delivered by humans. In contrast, switching to small, human-driven, electric delivery vehicles had a much more positive impact. “Going from internal combustion engine vehicles to electric delivery vehicles is really where the biggest impacts and opportunities lie,” Keoleian said.

But while automation doesn’t appear to impact emissions, Sam Heshmati, an industrial engineer at the University of Kentucky who was not involved with study, explained that there could still be indirect benefits. For example, with proper programming, companies could tap into the increased efficiency of autonomous systems to shorten delivery times . In a 2018 study, Heshmati showed that “even reducing one minute in [delivery] time had a huge impact on emissions.”

Also, even though the combination of self-driving van and delivery robot did not reduce the carbon footprint of home delivery, greenhouse gas emissions could theoretically be minimized if all-in-one delivery robots could do their jobs themselves. Many companies, such as Starship Technologies, are already exploring this route through the development of sidewalk delivery robots that can transport goods to customers up to a few miles away. This ultimately reduces the need for a roadworthy vehicle, removing these extra emissions from the equation.

But there’s still the matter of accessibility, since, much like e-scooters, delivery robots have the potential to make sidewalks more difficult to navigate, particularly for people with disabilities. This is something that Emily Ackerman, a postdoctoral researcher at Harvard Medical School and wheelchair user, experienced firsthand last year when a delivery robot blocked access to the cut in the curb of the sidewalk and almost left her stranded in a busy intersection on the University of Pittsburgh. At the time, Ackerman was a graduate student on campus.

“The crosswalk was ending, and I was ... trying to figure out how to get around [the robot],” she said. “I was put in danger by something that wasn’t sentient enough to understand that it was blocking me into the street.”

Eventually she managed to force her way past the robot and onto the sidewalk in a maneuver she has described as painful and dangerous. “I was truly upset. The curb cut was made for me, very specifically,” she said. Yet, her access was blocked by a robot that was simply following what it was programmed to do -- “wait for it to be safe for it to move.”

Later, she learned that other disabled people have had similar encounters with these delivery robots. “It was very much not a one-time thing, not even a one-company thing,” Ackerman said.

“We have to be smart about how we deploy technology, because we could have unintended consequences,” said Keoleian, adding that deployment needs to be strategic. “When does this automation make sense and when will there be benefits from it?”

Considering societal concerns is key to creating more impactful and sustainable delivery policies, added Heshmati. Otherwise, bad experiences with new delivery technology could postpone its use, regardless of its benefit to our environment.