If you grew up in the 70’s like I did, you surely have fond memories of watching Burt Reynolds as the Bandit hauling that big-rig down the highway, Smokey in hot pursuit. But is the Bandit about to be pink-slipped by R2D2?
The trucking industry is as old as it is massive. In the US alone, $700 billion in truck freight is driven by 3.5 million truck drivers working for 500,000 different trucking carriers. And the industry dates back to the dawn of the automobile. The Teamsters Union is named for the drivers of horse-drawn carriages after all. Many fleets are a family affair, which is not surprising given the low barriers to entry. More often than not, the founder myth starts “I bought my first truck used for $20,000…”
But would such an old-school and fragmented industry really adopt cutting-edge technology? In fact, it already has. Since deregulation in the ’80s, survival in the trucking business has been a game of intense competition, thin margins and constantly improving efficiency. Operating cost per mile has decreased by nearly 40% in that time, and only those companies that have kept on the forefront of efficiency have survived.
The 2000s were a period of focus on “hardware” — that is the truck itself. First through a switch from “long-nose” tractors to aerodynamic cabs, then through a series of changes in engines and emissions technology, a switch to automatic transmissions and now aerodynamic fairings on trailers, truck equipment makers have created significant fuel efficiency improvements over time. Since fuel is 30% of the operating cost of a truck fleet, these improvements are a big deal. Truckers may prefer the look of the Bandit’s truck, but they vote with their pocketbook wholeheartedly embracing the lower fuel costs of the new designs.
We have finally reached a point where hardware is delivering diminishing returns and now the game is all about software. Where the introduction of a next generation engine can bring 3–5% efficiency gains, the difference in fuel economy between a good and bad driver can be as much as 20%. Using software to measure and pay bonuses on driver efficiency and even provide real-time coaching for better driving behavior can bring real gains for fleets that can’t be had through improvements to the truck hardware itself. And given that some 20% of truck miles are currently being driven empty (as trucks are repositioned from one load to the next), using route and schedule optimization software to turn an empty mile into a loaded one brings not only more profits to the fleets, but lowers fuel consumption and emissions to boot.
So if fleets are looking to software to bring cost reductions and efficiency gains, isn’t it just a matter of time before trucks are driven by robots? The economics are certainly compelling. Driver pay is as big a cost as fuel. Worse, the industry has been plagued for a decade with a driver shortage and there is no end in sight.
In fact,The American Trucking Association projects a shortfall of 175,000 drivers by 2025 and this is driving a bidding war between fleets to hire and retain good drivers, further driving up labor cost.
The technology is certainly here. Google has famously proven their driverless technology over millions of miles on both highways and city streets, and Tesla has (infamously?) unleashed their Autopilot into consumer’s hands. The technology is already starting to make an appearance in trucks. Peloton Technology is using partially autonomous systems to create virtual “platoons” of trucks drafting like Tour de France riders to get 10% fuel efficiency gains. Meanwhile, Otto is going all the way and road-testing fully autonomous big-rigs on the roads today. Even the conservative European truck makers Daimler, Volvo and VW are road-testing autonomous trucks, albeit with somewhat less aggressive timelines than their Silicon Valley competitors.
Bandit vs R2D2
So what about that Bandit v. R2D2 deathmatch? I’m happy to predict that the robots aren’t going to take over any time soon. First we have the minor detail of regulation to contend with. The public is understandably nervous about an 80,000 pound truck barreling down the highway at 65 mph with nobody in the driver’s seat. But more relevant for regulation is the power of the trucking lobby.
The US Department of Transportation has been attempting to implement electronic logging of driver hours for the past 10 years — a seemingly beneficial technology — but has been stalled repeatedly by legal challenges from lobbyists representing the smaller players in the industry. These same smaller players are likely to be at a disadvantage relative to large fleets in their ability to purchase and manage driverless systems and so are likely to throw up similar objections to regulatory efforts promoting driverless technology.
Then there is the problem of loading. Driving a truck at a constant speed of 64 mph in a straight line down highway 80 might seem like an easy problem for a computer. But try figuring out which dock to pull up to, putting a tarp on a load, following unclear instructions over the phone from a freight broker, or negotiating an unloading price with “lumpers” when you are made of silicon. Drivers may lament that they have been reduced to “steering wheel holders” with the many restrictions in their jobs, but they serve a vital role in facilitating all of the human interactions on each end of the trip.
Much of today’s truck freight needs a driver, but there are places that a robo-truck can start to play a role. Cross-country routes of time-sensitive loads between distribution centers are a great place to start. With a driver in the cab to handle the pickup and delivery and to address issues that might arise on the trip (trucks do get flat tires), an autonomous truck could drive continuously from origin to destination while the driver slept during the uneventful middle of the trip — essentially like an airliner on autopilot. Today, high-value or time-sensitive loads like this are driven by pairs (“teams”) of drivers that trade off driving and sleeping. And as the public and regulators build up confidence in the safety of these systems on lower-risk routes, we could start to see ad hoc platoons of trucks with an on-duty driver up front and a sleeping driver or drivers behind.
Assuming that 10% of long-haul routes could take advantage of such a scheme which then doubles the miles driven per driver (by eliminating the 10.5 hours of rest they currently must take every 11 hours), the industry could add the equivalent of 180,000 drivers — just covering the anticipated driver shortage. Not exactly a robot takeover.
Longer-term the prospects of robot membership to the Teamsters do increase as facilities, systems and freight routes start to be designed for automation. And as the public and regulators become accustomed to summoning their driverless Uber/Google swarm-car in the cities, the notion of trucks without drivers will start to seem less terrifying. But just as robots have not completely supplanted the need for factory workers, the truck driver will continue to be a vital part of our labor force for years to come.
So perhaps our future looks more like a merger of those two classic 1977 films. The Bandit taking care of the human problem of evading Smokey while R2D2 as the trusty sidekick gets the Coors delivered on time.