TuSimple is part of a stable of startups seeking to automate parts of long-haul and parcel transport with self-driving technology that uses artificial intelligence, laser sensors and cameras to navigate roads. TuSimple, Embark, Starsky Robotics and truck-platooning business Peloton Technology Inc. have drawn strong interest from venture capital funds and have struck a variety of operating agreements with truck manufacturers and operators to test their technology.
TuSimple’s technology is built around cameras that it claims provide better long-range predictive capabilities than lidar, the technology used in most self-driving passenger cars that offers a 3-D laser view of the environment.
The company’s cameras can see about 1,000 meters, or 3,280 feet ahead, said Chuck Price, TuSimple’s chief product officer. “From a half mile away we can spot emergency vehicles, cars broken down on the side of the road, people walking around,” Mr. Price said.
The company has two delivery routes in Arizona that deploy its technology on retrofitted trucks, with backup drivers and engineers on board, that haul loads for a dozen customers that it declined to name. The average run is about 200 miles and is automated from end to end, including surface-street navigation, Mr. Price said, although the trucks need a human driver to back up to loading docks.
TuSimple is working with truck manufacturers Navistar International Corp. and Paccar Inc. and components suppliers such as engine-maker Cummins Inc. The company will use the new investment to fund joint development with those companies to integrate autonomous software with powertrain, braking and steering systems as it pushes to achieve commercial scale.
“We are confident that we will have our first commercial driverless operation in late 2020 to 2021,” Mr. Price said.
- Starsky Robotics
- Peloton Technology Inc
Of those companies, I am confident in Embark, TuSimple, Waymo, and Starsky Robotics.
Spotlight Starsky Robotics
In March of 2018, Wired reported Starsky Robotics Unleashes it Truly Driverless Truck in Florida.
Starsky doesn’t want humans in truck cabs at all. “We want to get people out of the cab because the work is unpleasant and dangerous,” Seltz-Axmacher says. Today’s trucking work, he argues, is bad, with uncomfortable work environments, long hours that leave little time for friends and family, and wages that aren't high enough to compensate for those downsides. That’s why annual driver turnover in large American fleets hit 95 percent in 2017, according to the American Trucking Associations.
Like Uber and Embark, Starsky’s trucks will handle the highway driving all on their own. But when a human grabs the wheel to negotiate the complex surface streets, they won’t climb into the cab to do it. They’ll work in buildings that look like call centers, monitoring 10 to 30 vehicles per hour via video links and using a videogame-controller-like wheel to take control as needed. (Today, the company employs four truck drivers.)
Which model of robo-trucking the future embraces is probably up to regulators as much as the free market. (Starsky, for its part, just announced a $16.5 million Series A funding round, led by Shasta Ventures.) Today, eight states permit trucks to “platoon”—that is, use sensor integrations and wireless communications to synchronize accelerating and braking between two or more vehicles, so that only one driver (the one in front) has to pay attention at a time. Peloton Technologies, a California company that has embraced platooning, says it will begin to make commercial deliveries this year.
Which Model Will Win?
I expect numerous models will win starting with Embark.
On February 3, 2019, I commented Amazon Hauling Cargo on I-10 in Self-Driving Trucks Developed by Embark.
Embark, Electrolux, and Ryder have partnered in a driverless truck endeavor on a 650-mile I-10 route.
Embark CEO Alex Rodrigues: "Embark, Electrolux, and Ryder are working together running the longest automate freight rout in the world. 650 miles starting in Texas and ending in California. On the Frigidaire line, we drive over 100 million miles a year."
"Embarks approach is unique. Our automation is designed specifically for the highway. We rely on Ryder's trucks and drivers to ferry freight between the warehouse and the interchange.
Embark's trucks pick up at the edge of the interstate and from there, the computer drives it 650 miles, all the way to California."
That is the model I envisioned a decade ago. Platooning can easily be a part of that model. And 8 states already allow it. I expect the US Department of Transportation (DOT) will soon mandate allowance in all 52 states.
Starsky’s model goes even further. It completely solves the "last mile" problem.
This level of competition guarantees one or more self-driving models will be successful.
There is no other realistic way of looking at this.
The timeline primarily now depends on approval from the Department of Transportation (DOT). It will come withing a few years, most likely two.
Mike "Mish" Shedlock