Elon Musk’s grand plan of moving beyond passenger cars to truly revolutionize transportation just got a bit grander. In addition to developing an electric 18-wheeler that Tesla plans to unveil next month, Musk wants to make the thing drive itself.
Tesla is working with Nevada authorities to begin testing a robo-rig prototype at some point in the not-too-distant future. “Our primary goal is the ability to operate our prototype test trucks in a continuous manner across the state line and within the States of Nevada and California in a platooning and/or Autonomous mode without having a person in the vehicle,” Tesla’s Nasser Zamani told officials with the Nevada Department of Motor Vehicles, according to Reuters.
Assuming Tesla can figure out how to make battery tech work for long-haul trucking (no easy feat), adding autonomy to the equation makes perfect sense. Tesla joins a long list of enterprises working on autonomous long-haul trucking, including Uber, Google spinoff Waymo, Volvo, Daimler, the US Army, and a small horde of startups.
They all see a compelling case for human-free trucking: Big rigs carry 70 percent of all goods shipped across the US, but the industry doesn’t have nearly enough drivers. The American Trucking Associations says the industry needs another 50,000 drivers, and that figure could hit 175,000 by 2024 as more people retire or move on to other careers.
Autonomy also could drive down costs. The American Transportation Research Institute estimates that driver pay and benefits account for nearly 40 percent of a shipping company’s costs. And don’t forget safety. Big-rig crashes kill 4,000 people on US roads each year and injure another 116,000. Nearly all of them are the result of human error.
The great news is that the technological challenge of making a truck drive itself on the highway is relatively simple. Compared to navigating city streets swarming with cyclists, pedestrians, and unpredictable motorists, keeping any vehicle within its lane on the open road and maintaining a safe following distance is a breeze. The cost of all those sensors—lidar and top-shelf GPS can run five figures—is a bit easier to swallow in a big rig that costs 120 grand.
But—and there’s always a but when you’re dealing with leading-edge tech—the technology doesn’t really solve the driver shortage problem or truly pay for itself until you kick the human out of the cab entirely. And that is where Musk’s plan faces its biggest challenge.
Ditching the carbon-based life form altogether remains tricky, because although trucks spend most of their time on the interstate, they do venture into trickier terrain: surface streets, shipping ports, cargo terminals, and the like. And don’t forget the other tasks drivers tend to: hooking up trailers, checking tires, pumping fuel. Try to develop the technology to handle all that stuff, and autonomous trucks go from relatively easily to pretty damned hard.
And so the clever folks hammering away at this problem have a few workarounds Musk & Co. can pursue.
This is also known as a road train. The idea is to wirelessly link vehicles, so each follows the leader like railroad cars following the locomotive. Even with a human driving the lead truck, this cuts labor costs significantly and increases fuel economy. Running vehicles that closely together reduces wind resistance, much like a peloton of cyclists in the Tour de France.
The idea has been knocking around for years; Volvo trialed it in Sweden back in 2012. Tesla is pursuing it too, according to that Reuters report.
The big challenge here is motorists like you, who may freak out seeing a road train hundreds of feet long bearing down on them. Passing a platoon or waiting for it to clear that exit or on-ramp might annoy people. Oh, and the best way of having the trucks in a train communicate remains TBD. And what happens when it’s time to get off the highway onto city streets?
Don’t like trains? How about boats? In the tugboat model, a human drives the truck from the terminal or depot to a staging area on the highway, then turn things over to the computer. The same thing happens at the other end: A driver meets the truck at the highway and drives to the final destination.
This setup, where drivers function like tugboat pilots (actually, bar pilots, but no one knows what they are, so just go with it) guiding container ships into port, minimizes the number of humans required and eliminates the tedium of long days on the road.
“That’s how you give drivers a better quality of life,” says Alex Rodrigues, the CEO of Embark, a startup pursuing what he calls an “exit to exit, highway self-driving truck.”
Pulling this off requires an autonomous system so robust it never breaks down or freaks out. And the details on those staging areas—where they are, who can use them, how they’re marked and designed—remain TBD.
If you want a human in control but not in the vehicle, take the approach pursued by Starsky Robotics. The San Francisco startup (pronounced like Starsky and Hutch, not star-sky) sees a world where trucks drive themselves on a highway, but a human in remote location takes over when needed. Drones offer an analogy here: the “driver” sits in a simulator, working the wheel and pedals remotely.
“We’ve cut this really hard problem down into a very digestible chunk,” says Starsky CEO Stefan Seltz-Axmacher.
Obviously, the hurdle here is ensuring that the latency between the remote operator issuing a command and the truck responding to it must be nearly nil.
OK, maybe this problem isn’t quite easy after all. But that’s never stopped Musk before.