Editor’s note: This is an edited version of a presentation made by Trucks.com Editor Jerry Hirsch at the Autonomous Vehicles Silicon Valley conference on Tuesday.
When we think about the advancement of self-driving vehicles it should start with trucking, freight and logistics.
Trucks move about $700 billion of cargo in the U.S. annually, which amounts to about 70 percent of the nation’s freight by weight. In the near term, that’s why the economic case for driverless technology is so much greater for commercial vehicles.
While many look ahead to the decline of human driving and a subsequent decrease in crashes and traffic fatalities and injuries, most agree it remains a distant vision.
Just last week the Center for Automotive Research in Ann Arbor, Mich., said that self-driving vehicles that fit the Society of Automotive Engineers’ Level 4 and Level 5 autonomous operation rankings won’t even reach 4 percent of new-vehicle sales by 2030. Beyond that, it’s possible they could reach 55 percent by 2040.
As of last year, barely half the nation — 26 states plus the District of Columbia — started to craft regulation for automated vehicles, according to the research group. That leaves us asking how will we get there, what will it take?
Human transport is dividing into two forms: lifestyle mobility and commodity mobility.
Lifestyle mobility is where individual vehicle ownership and driving will remain. People will own vehicles that fulfill a lifestyle and facilitate activities like skiing, biking, kayaking or boating.
Commodity mobility will encompass much of human transport. This could be walking or biking to work, taking public transit, Uber or driving a basic Toyota Corolla. Commodity mobility is where we see the roll out of autonomous vehicles. But it will happen first in endeavors where there’s an economic payback for the operator, such as freight and logistics.
If you’re a fan of comic books or science fiction you will recognize this photo from a scene in the “X-Men” movie franchise Logan. The movie takes places in 2029 and opens with a scene of Logan driving a futuristic Chrysler limousine.
This is an interesting thought. If you ask Waymo or Uber, I think they would say limos will be autonomous more than a decade from now. Maybe, or maybe there will be a market for transport in unconnected cars, vehicles that are harder to track in real time, used by people who don’t want others knowing where they are travelling.
Later in the movie Logan is out on the highway helping a family corral horses that got loose from a trailer. What he calls “autotrucks” whiz by at high speeds. These look like strings of two to three large shipping containers that drive themselves. There’s no cab, no driver, no human involvement. Everyone here would see these autotrucks as full Level 5 capable of driving at well into triple-digit speeds.
Despite the features automakers and technology developers are now packing into light vehicles, the true rollout of autonomous driving will be gradual and start with commercial vehicles. There’s a pretty simple reason. It’s where it makes economic sense. Think about it this way. Velodyne recently slashed the price of its popular lidar unit, a key sensor needed for autonomous driving, in half to $4,000.
Consider this: The top-selling vehicle in the U.S. is the Ford F-Series pickup truck. It starts around $35,000 for the crew cab version. The best-selling passenger car that isn’t a truck is the Toyota RAV4, which starts at about $25,000.
Following Velodyne’s price cut, the cost of the lidar unit would add at least 11 percent to the price of the Ford truck and 16 percent to the RAV4. How many consumers are going to pay that premium to get new, unfamiliar technology to perform a task they believe they can do perfectly well right now? Not that many at the outset.
A fleet manager for a trucking and freight company will see that differently. Fleet managers will plug that extra expense into a spreadsheet. The manager will evaluate what efficiencies can be gained from the technology and calculate a payback. When that payback drops to a certain level— boom! — it’s time to buy.
This creates a powerful economic incentive to transition to self-driving trucks. As volumes increase, the price for the sensors and related equipment and software drop, shortening the payback and hastening the transition. Eventually manufacturing economies of scale drop the cost so the tech can penetrate the consumer market.
The transition will also take place in the last-mile delivery sector. IHS Markit estimates e-commerce made up 16 percent of U.S. retail sales in 2017, up from about 6 percent in 2007.
That torrid growth is expected to continue.
Just imagine all the delivery vans and drivers needed to satisfy demand for package delivery. Already companies are looking to solve this problem with autonomous vehicles. These are purpose-built vans loaded with fashions from shops at the local mall, electronics from e-retailers around the globe and groceries from neighborhood purveyors.
Just last in San Mateo, autonomy startup Udelv used a custom-built package delivery van, bristling with rooftop cameras and lidar sensors, to make a roughly 3-mile run on public streets to deliver groceries from Draeger’s Market. Waymo, Ford, Daimler and others are working on automating last-mile delivery.
Once you figure the transition will start in the commercial sector you have to ask where.
The sexy part of the business right now is in Robo-Taxis. This is the primary direction Waymo is headed, testing a service in Phoenix. GM also recently said it plans to launch an autonomous taxi service next year.
Once you take the driver out of the car the amount you can charge per seat mile and still be profitable plunges. Here are the numbers to keep in mind. In Los Angeles the cost per mile in Uber is about $1.55. The IRS pegs the cost of individual vehicle ownership at about 54 cents a mile. The faster a Robo-Taxi service can beat those numbers, the faster you can scale up.
But this is an all-or-nothing approach — either you have a driver or you don’t. I don’t see the widespread deployment of autonomous technology developing that way. These services will be experimental and only in limited urban areas.
Freight is where the real action will take place. It will start with what I am calling a driverless creep. Don’t expect to see Logan-style drone trucks plying the highways anytime soon.
As the nation’s heavy-duty truck fleet turns over, each new model year vehicle will be equipped with slightly more autonomous technology. The concept is to make driving for truckers easier, improve safety and increase efficiency.
The truck fleet is smaller and turns over at twice the pace of the consumer car park. That’s why this is where the transition will take place.
There is a key regulation we have to watch even though it says nothing about driverless technology.
It is the much-debated federal hours-of-service rule. The regulation limits truckers to driving no more than 11 hours a day within a 14-hour workday. Drivers must then be off duty for 10 consecutive hours. It’s a little different in Canada where truckers are held to driving no more than 13 hours a day within a 16-hour workday. Drivers must then be off duty for eight consecutive hours.
Think about the economic impact of automated driver assistance systems that would expand the U.S. trucker workday to that of Canada.
The trucking industry, which currently has a direct pipeline into the White House, will push hard for loosening the regulation if it can prove that autonomous technology can improve safety and allow for a longer driving day. The big U.S. motor carriers purchase trucks by the thousands annually.
They would rapidly transition to trucks that meet the technology test to allow for longer hours of operation.
There are other ways that autonomous driving technology can improve efficiency.
Think of the platooning concept where digitally tethered strings of two to five trucks travel closely together to decrease drag and improve fuel economy.
Think of hands-free, feet-free driving technology akin to what Tesla calls Autopilot. Let’s say a trucker can get on a long highway and set the vehicle to drive itself. The robot will get better fuel efficiency by calculating when to accelerate, brake or shift. Predictive cruise control can automatically look at the varying grades of the next section of road and decide where to apply acceleration and where to coast to get maximum fuel economy.
Think about trucks that go into drone mode once they hit the highway. The driver is just there for controlling first- and last-mile driving. You don’t have to program the truck to drive through downtown Chicago. The operator could be sleeping through most of the trip and be fresh to grab the next load after delivery.
Farther out, think of trucks programmed to go on a particular route, such as the I-5 from Bakersfield to Sacramento. A truck that can do that without a driver would be extremely valuable. You might have a driver hop on or off at a central depot outside the urban area to guide the truck to its end destination.
We will see two complementary transitions. One will be that of a truck to an autonomous device. The second will see the traditional truck driver change to a load manager or freight conductor.
I hear this idea that 3.5 million truck driving jobs in the U.S. are on the cusp of disappearing. To this assumption, I just shake my head. Yes, their jobs will change but not fully disappear.
Also, that figure, which is used by the American Trucking Associations, is an overestimate. There aren’t that many drivers in the U.S. According to Trucks.com research, there are about 2 million drivers employed full time hauling goods, a calculation based on Bureau of Labor Statistics data.
There’s an abundance of trucking jobs in the near term but no one to fill them. That’s because there aren’t many people who want to do the job, and almost no young people. Automation can help plug the gap, creating yet another economic incentive for the industry to push self-driving technology that can increase productivity.
In truth, we are facing a technological transition where we don’t know the time frame or what it will do to the marketplace. This will all sort out in the decade leading up to 2030.
But if you are looking for hints in how it will play out, focus on who has the biggest economic incentives to move to driverless technology. It’s not the pod farm occupant driving to work or the parent taking the children to soccer, it’s the people that haul the goods from our ports, farms, factories and e-commerce centers.