Are autonomous trucks the future of freight transportation? A number of the finest minds in automotive engineering are convinced the answer is yes, but they may have a long road ahead of them.
“Autonomous trucks are coming,” said Mike Roeth, the executive director of the North American Council for Freight Efficiency (NACFE). He says the only questions are when they will begin arriving as production vehicles, how deeply integrated the technology will be into the vehicle and what autonomy level they will have.
True autonomy, as defined by the Society of Automotive Engineers’ scale as Level 5, where vehicles operate full time, in any conditions, without a human pilot, is still an ambitious goal. Numerous challenges remain, principally the technology’s ability to deal with the unpredictable nature of real-world driving.
With its support for innovation through tax credits and research funding, as well as regulations that allow for autonomous vehicle (AV) testing, Ontario has attracted numerous startups that are tackling the engineering problems that come with autonomous vehicles.
Gatik AI Inc. and NuPort Robotics Inc. are two startups taking advantage of the supportive environment. While they are taking different approaches, both believe the “middle mile” will be the sweet spot for trucking.
The middle mile describes that part of a product’s journey between the manufacturing site and a distribution centre, or between a distribution centre and a retail outlet. It is distinct from the “final mile,” or “last mile,” which delivers to an end customer’s door. Gatik and NuPort Robotics are tackling middle-mile autonomous deliveries over relatively short distances, no more than 40 kilometres.
The middle mile is appealing for these companies because they say it offers the potential for a more predictable, less chaotic environment. And chaos is the enemy of autonomous driving. The principal problems lie in what engineers call “edge cases.” These are situations that are completely unexpected – such as a unique sign or unforeseen kind of wildlife on the road, for instance – but that the autonomous system must be able to deal with. The more predictable the environment, the fewer edge cases should be encountered.
According to Gatik’s head of policy and communications, Richard Steiner, operating over the middle mile improves an autonomous vehicle’s chances of success by operating on “fixed, repeatable routes.”
Gatik has been making middle-mile deliveries using driver-assisted autonomous trucks for grocery giant Loblaw Cos. Ltd. since 2020. Ferrying goods from the distribution centre in the Greater Toronto Area to local stores on the same routes day in and day out helps to eliminate unpredictability, Mr. Steiner said.
Gatik, which is based in Palo Alto, Calif., and Toronto, has not yet removed the human driver from the equation. However, Mr. Steiner says the company has received approval to do driverless trials in Arkansas, where it works with Walmart, and these tests will go ahead this year.
Toronto-based NuPort Robotics is working with Canadian Tire on a similar project, moving goods in a 20-km radius from rail depots to the distribution centre in the GTA. NuPort chief executive officer Raghavender Sahdev says that by staying out of the city core, they have reduced the complexity of the route.
“Humans are very unpredictable and it’s tough to have autonomous driving where there is a lot of traffic,” Mr. Sahdev said. “I’m not saying it’s impossible; it’s just much easier when there is less traffic.”
Raquel Urtasun is the founder of autonomous-trucking startup Waabi Innovation Inc., which also makes its home in Toronto. While her model focuses on long-haul trucking, because she believes highways are less complex and will yield faster and safer results than either Gatik’s urban roads or NuPort’s suburban byways, she shares their optimism that AVs are coming.
“I truly believe this is going to happen, and my life’s mission is to provide self driving to the world,” she said. “It’s going to happen; it’s a matter of time.”
Autonomous trucking will offer many benefits, proponents say. Mr. Steiner says Gatik’s middle-mile solution offers reduced costs and faster turnaround times, while NuPort’s Mr. Sahdev argues that greater efficiency will mean fewer trucks on the road, which will improve the carbon footprint of trucking and help alleviate the shortage of truck drivers. Mr. Sahdev notes that greater road safety will be another benefit.
Not everyone agrees. Francesco Biondi, an assistant professor in the department of human kinesiology at the University of Windsor, says the belief that autonomous driving will improve highway safety is based on a myth. If 95 per cent of accidents are caused by human error, and you remove humans from the equation, the theory goes, then accidents won’t happen.
“I think there is a bias that humans are dumber than a machine, or we’re just dumb,” he said. “The reality is that these machines may be good at completing individual tasks, but in terms of the ability to integrate information, make decisions using the information that they perceive, and execute behaviours and actions that are consistent with the decision making, there’s no competition [with the human brain] today.”
While Biondi, who is 34, says he doesn’t believe he will see widespread adoption of autonomous driving in his lifetime, he does believe that if AVs can be segregated from general traffic, they will act like a train. He said the model that employs a human driver to pilot the truck from the warehouse to the highway and then from the highway to the final destination is most likely to succeed.
“They won’t have a railroad, but they won’t encounter other cars or road users cutting them off or being in their path,” he said. “It’s a train without a railroad, and we know trains are safe.”