Just over two years since UPS made a minority investment in TuSimple, the self-driving truck company and the largest U.S. carrier accumulated more than 160,000 autonomous miles and found significant fuel savings, TuSimple announced in a Nov. 3 letter to shareholders.
UPS made that investment after starting a pilot program with TuSimple that has since expanded to a nearly coast-to-coast autonomous freight network from Arizona to Florida. The two companies mapped new autonomous lanes between Dallas and Orlando and up the East Coast to Charlotte, North Carolina.
Moving actual freight has helped TuSimple grow its offerings and develop real-world solutions for the future of transportation, according to CEO Cheng Lu. “To actually enable and leverage technology to enable freight capacity, you have to have a real-life application of it,” he told FleetOwner. “And we have to work with real-life customers to understand their needs.”
Those needs, he said, go beyond self-driving technology to include “a level of service. UPS has been very forward-thinking. Everyone recognizes that there’s a supply chain issue, there’s a driver shortage issue. So how do you integrate more new technologies into their network? Since Day One, we’ve been busy hauling freight for them using our fleet of prototype, retrofitted autonomous trucks.”
While TuSimple gets paid for hauling the freight—specifically for UPS’s North American Air Freight (NAAF) division—the “testing miles and data collection” has helped TuSimple improve its technology, Lu said. “For them, we share many different metrics so they can understand how autonomy kind of evolves.”
TuSimple is collaborating with UPS Supply Chain Solutions to expand its autonomous freight network “ahead of schedule to the East Coast to reach UPS NAAF terminals in Orlando and Charlotte,” according to the letter to shareholders. “We have already expanded our network of HD mapped routes to Orlando and Charlotte, readying us to commence autonomous commercial runs to those locations.”
Big fuel savings
Lu called the shipper-autonomous driving company relationship’s 160,000 miles together “a big deal.” As they expand their routes together, the milestone not only “validates how well we’re doing in the program, but there’s a capacity crunch. Anytime you can provide some additional help to alleviate that supply chain, everyone’s very welcoming of it.”
Beyond the supply chain benefits, Lu said what he finds most compelling is the 13% fuel savings that TuSimple’s autonomous vehicles achieved on long-haul operations while traveling 55 mph to 68 mph, which is also the highest speed UPS allows its human-driven tractors to travel. “The fuel efficiency comes from just the behavior of the autonomous driving system, which has a lot more consistent acceleration and deceleration curve,” he explained.
Those trucks achieving these fuel-economy successes have safer operators on board for now. But the trucks are run by artificial intelligence, which plans further ahead than human drivers. “Our trucks don’t have harsh braking, they’re not switching gears as much, and all these other things that kind of burn a little more gas,” Lu said. “Of course, a lot of people in the (AV) industry have talked about how their systems can save 10-plus percent on fuel economy. But for us to have empirical data—validated by a major carrier—and to have 13% fuel efficiency, I’m really proud of that.”
AVs’ fuel economy benefits are often overlooked, Lu noted, because the “headline news” focuses on removing the driver. “But to put it in perspective: This is an $800 billion industry, and fuel is the second-largest cost item—it’s about 25 to 30% of the cost per mile. For an industry that operates on the margins, this translates into double-digit billion dollars of savings. That alone is really a testament to how this technology can be helpful to the supply chain in general.”
The mapping
TuSimple’s approach to autonomous freight movement is catching the attention of experts across the industry. “TuSimple has driven tens of thousands of miles without their system failing and requiring the driver to take over—that says something,” Criss Wilson, a data scientist at McLeod Software, said during the McLeod User Conference in September.
“That says they’re figuring it out,” Wilson said, crediting the AV company’s multilayered mapping for keeping its vehicles safe even if road conditions change because of construction or other reasons.”It's in that geometry layer, it's in the semantic layer, and it's in the priors layer. It's that triangulation of those layers (that tell the vehicle) where you are and where you should be.”
Lu said TuSimple can create new route maps “in under two weeks. We have smaller mapping vehicles that go and collect the information. That creates the 3D, multilayered map. The benefit of the map is it provides an additional layer of safety for the operations.”
Those additional layers, which could include landmarks along roadways, help TuSimple’s trucks know where they are. Lu said that his company is seeing people “kick over construction cones” along autonomous freight routes.
“I don’t know if it’s an internet meme or something—but people are actually really pushing over construction cones,” he explained. “So our cameras can detect construction cones—but the ones that fall over can create confusion and little more risk. But having updated maps allows us to have another way to triangulate to make sure that we’re operating safely.”
While Lu said that’s a small example, it shows the importance of mapping and continuously updating those maps “because of construction, changes to the landmarks, and so forth.”
The TuSimple CEO stressed that his company is not in “the business of selling trucks. We’re in the business of automating lanes. And not every lane makes sense to be automated.”
Lu wants to continue to focus on high-volume, long-distance freight routes. “These are also the lanes that probably require team driving and [are] harder to fill,” he said. That will also include scaling out AV support operations, such as refueling and roadside maintenance services for more remote highway areas.
“So you can imagine us kind of scaling out in an isolated geography rather than having a truck all of a sudden that can drive anywhere in the U.S.,” Lu explained. “That’s why it's so important to work with carriers and shippers today because they need to understand as they're planning ahead what this technology can do and cannot do to fit into the supply chain.”
