A Georgia Tech study funded by Ryder System Inc. suggests huge cost savings for autonomous trucks using a network of autonomous transfer hubs that gain efficiencies beyond avoiding the cost of pay and benefits for human drivers.
Ryder (NYSE: R) provided order data from its dedicated transportation solutions operation in the Southeast to researchers at the H. Milton Stewart School of Industrial and Systems Engineering at Georgia Tech in Atlanta. Miami-based Ryder has alliances with four autonomous trucking startups — TuSimple, Waymo Via, Embark Trucks and Gatik.
“Since we’re putting so much figurative investment into the space, it was really important for us to start to build some of our own knowledge,” Mike Plasencia, Ryder executive director of innovation, told FreightWaves. “We’ll never build the truck … but there’s a lot of other places we want to explore and play a role in this ecosystem.”
Transfer hub networks would use autonomous truck ports to hand off trailers between human-driven trucks and driverless autonomous trucks. Autonomous trucks would carry freight between the hubs. Conventional trucks with human drivers cover the first and last miles.
Operational cost savings could range between 22% and 40% in the transfer hub model based on the cost difference between driverless trucks and conventional trucks, according to a separate study by management consulting firm Roland Berger.
“The assumption we used in the [Georgia Tech] study was that when an autonomous truck is running, it is 25% less than the cost of the truck with a driver,” Plasencia said. “Some of that could be pay. Some could be fuel. We focused more on what was the network opportunity.”
The Georgia Tech study estimated operational cost may be reduced by 27% to 38% for a small network and 29% to 40% for a more extensive network. Estimated savings of $5.5 million to $8.4 million per year were based on the orders Ryder contributed to the study.
“I’ve worked on a lot of different transportation problems in the past, and if you have 1% improvement that’s magic,” said Pascal Van Hentenryck, a Georgia Tech professor and associate chair for innovation and entrepreneurship. “Here we’re talking about 29-40%, so it’s massive. It’s really massive.”
The hubs apply automation on highway stretches while more complex local driving and customer contact work remains with humans. Fully robotic trucking is expected to focus on long-haul freight, a segment in which driver turnover is approaching 100% a year and the driver shortage is pronounced. The first driver-out trucks are expected in late 2023 and 2024.
Increasing asset utilization
In dedicated transportation, the focus of the study, the opportunity to increase asset utilization comes by negotiating appointment times with customers. In today’s model, orders are typically placed multiple days in advance.
In an autonomous network, trucks can immediately start serving the next order without returning empty. Autonomous trucks also do not need to return to a base at the end of the day and can run around the clock except for maintenance. Tractor and trailer utilization can double the hours of service regulation that limits human drivers to 11 hours driving in a 24-hour period.
In the transfer hub model, vehicles still transport one order at a time, but the one-to-one matching between a customer and a truck/driver is eliminated. Human drivers would exclusively serve the first- and last-mile legs for multiple orders on the same day.
For example, a driver may bring freight from a transfer hub to a last-mile customer, drive empty to the next customer to pick up a new load and deliver it to the first-mile hub. Variations of this approach would play out throughout the day, allowing the driver to go home at night.
‘You’ve got to introduce flexibility’
Autonomous transfer hubs would reshape the traditional business model in significant ways, the study found. One is knowing the parts of a run when the truck is moving freight and when it is driving empty, allowing more efficient use of equipment.
The timing of the orders is key. In the Ryder case study, the timing was based on original appointments. Making appointments flexible by 30 to 90 minutes could add efficiency. Detention, which eats into a driver’s allowable time behind the wheel by delaying loading or unloading, was not included in the study.
“You’ve got to introduce flexibility,” Plasencia said. “When autonomous trucks come, if shippers want to be able to get the value of that capacity, they’re going to need to be more flexible. Having the autonomous truck always be on this rigid schedule is really going to limit their ability to deliver the value.”
The case study showed that autonomous transfer hub networks can operate with comparatively few autonomous trucks because of their ability to operate day and night. The number of fully robotic trucks is expected to remain in the thousands or tens of thousands in the next decade, a fraction of the 2.5 million heavy-duty trucks in use.
Most autonomous truck testing is occurring in the Southwest, though TuSimple (NASDAQ: TSP) plans to operate human-supervised trucks in Florida and North Carolina by the end of the year. Ryder wants “real alliances” with all four autonomous trucking software startups with which it works.
“We’ve had a chance to get to know these companies and what they’re doing,” Plasencia said. “We believe autonomous technology is coming. There’s going to be multiple players in the space and we want to provide the best solution, not only for our customers, but for these autonomous companies.”