Study Finds TuSimple Trucks At Least 10% More Fuel Efficient Than Traditional Trucks

Autonomous trucking companies have long argued that self-driving technology will not only make trucking safer and more cost efficient but that it will also help reduce the amount of pollution commercial vehicles emit.

Now a University of California San Diego study has substantiated some of those claims, with findings showing that autonomous trucks operated by self-driving startup TuSimple reduce fuel consumption of heavy-duty trucks by at least 10% and up to 20%.

"We were surprised by the data," Henrik Christensen, director of the UC San Diego Contextual Robotics Institute and computer science professor, told FreightWaves. 

The research team expected to see some savings, Christensen said. "But 10%, we were like: ‘Wow, this is a big differentiator.'"

The study, which compared the fuel efficiency of TuSimple's Level 4 autonomous trucks with the efficiency of trucks manually operated by a driver, comes with some qualifications. The most significant fuel-efficiency gains occurred in trucks traveling around 10 miles per hour and declined as speeds increased.

"When you are coasting on an open highway with no traffic around you, the truck driver and Level 4 truck get basically the same performance. The big difference is when you drive in traffic."

To conduct the study, the UC researchers fitted the trucks with a black box that extracted driving data over a period of six months, which included 122 autonomous missions totaling nearly 6,700 miles. 

Based on the available data, the team used a Virginia Tech fuel consumption model to estimate fuel consumption as a function of speed, location, acceleration and braking.

Manually driven trucks were fitted with the same black box technology. Both sets of trucks were hauling cargo for corporate customers on routes in Arizona. 

The research team then analyzed fuel consumption across different ranges of speed: 0-30 mph, 30-40 mph, 40-50 mph, 50-60 mph and 60 mph or more. 

Trucks in the first category realized up to 21% efficiency gains compared to manually operated trucks, while trucks in the over 60 mph category saw under 3% gains.

Without data from other autonomous trucking companies, Christensen said he could not comment on extrapolating the findings to other self-driving truck startups. 

But he strongly suggested that the TuSimple study was good news for the industry overall. TuSimple uses Peterbilt trucks with standard engines, he said, "good but standard trucks."

The study did call attention to different autonomous trucking use cases — and where autonomy might provide the most bang for the buck. There may be relatively small efficiency improvements to be gained by automating free-flow highway driving, but most self-driving startups have focused on long-haul driving, reserving drivers for complex urban driving tasks. 

Not all freeway driving is a cruise on the open road, of course. The stretch of Interstate 5 running through San Diego is "a parking lot" half the day, Christensen observed. "If you can save 10%, even 20%, that's a pretty big deal for the trucking companies. We're talking real money."

TuSimple said in a statement that applying a 10% fuel savings to all U.S. diesel trucks would result in an estimated savings of 4 billion gallons of fuel totaling $10 billion per year, according to calculations based on ATA fuel consumption reports.

If all medium-duty and heavy-duty trucks adopted TuSimple's self-driving technology, according to the statement, it would result in an estimated savings of 42 million metric tons of CO2 emissions per year, based on calculations using the EPA's Inventory of U.S. Greenhouse Gas Emissions and Sinks report that outlines transportation emissions.

"The research findings from UC San Diego validate our belief that TuSimple trucks will significantly reduce the environmental impact of heavy-duty trucking," Arda Kurt, director of motion planning and control at TuSimple, said in the statement.

"We plan to optimize our technology for fuel and emissions savings in the future, meaning these numbers will be even more significant as we further develop our technology."

Image by MikesPhotos from Pixabay