What If It Works? The Autonomous Truck Revolution

The tech, economics, and second-order effects of driverless trucking.

Iconic Truck from Convoy, a 1978 Movie

Trucks have a nostalgic quality to them: pumping my fist in the backseat as a kid to get a horn honk was about as euphoric as it came for a 6 year old. There was something undeniably “cool” about the American trucks too - as I grew up in the UK with the flat nose type which just lacked the visual appeal.

That nostalgia for that time will only grow because in the march towards automating everything, trucks are the next domino to fall, and a lot will change as a result. In fact, the future is already here: Aurora has driverless trucks on the road in Texas already. No more horn honking for the kids.

The Law of Freight: Time vs Money

Before we get into the fun of autonomous trucks and how the future looks with them as part of it, let’s talk about why trucks are America's vehicle of choice for freight. First and foremost, you have to always keep in mind that freight is always a tradeoff between cost and speed. To get freight over an ocean for example, you can do it fast and expensive in a plane, or slow and cheap in a boat.

Over land, the equivalent is: fast and expensive on a plane, slow and cheap on a train, or medium fast and medium expensive on a truck. Occupying that medium speed/cost slot is what makes trucks so popular. The door to door nature of trucking is a big part of what makes it appealing too. I’ve experienced this first hand when I’ve shipped products from a factory to a warehouse for the business I founded. Often if we don’t quite fill a whole truck, we will still book the whole thing so that we can get door to door service with no stops and no other inventory being loaded onto the truck for possible cross contamination. There’s a psychological safety to knowing your product is just going door to door.

So how does this tradeoff change once trucking goes fully automated?

It remains somewhere between planes and trains, but gets a little cheaper and a little faster, making it even more competitive. Here’s how:

24/7 Trucking

Today the Department of Transportation limits the hours truck drivers can operate for the purpose of safety. A great idea given that a sleepy driver of an 18 wheeler can do a lot of damage.

The rules are as follows:

1. 14 hour shifts, only 11 hours of driving per day within that shift window.

2. After 8 hrs of driving the driver must take a 30 minute break.

3. No more than 60 hours of driving in 7 days, and no more than 70 hours of driving in 8 days. To reset this rolling clock you have to take a 34 hour break.

So in a 7 day period, assuming a driver hits their maximum limit, they are driving 60 hours out of the 168 available. That’s a roughly ~36% utilization rate of the truck.

Now some truckers have a sleeper cab and trade off shifts with another driver. This doubles their time on the road to get a utilization rate of ~72%. This arrangement is less common, and typically only on the long haul routes where it's advantageous to drive through the night.

Bring in a driverless truck, and none of these rules apply. The driving system doesn’t get tired. Fuel, maintenance and loading/unloading are now the only limiters to how many hours a week a truck can drive. So deliveries get faster, and without having to pay drivers, cheaper too.

So if freight is now 24/7, faster, and cheaper than before: what else changes?

24/7 Warehouses:

Trucks are reliant on the nodes they go between. Picking up and dropping off the freight. If we unleash the trucks to be fully autonomous, so they can maximize the utilization rate of the truck, we might want to consider keeping our warehouses running 24/7 as well. Warehouse automation then becomes an imperative for businesses that want to take advantage of the 24/7 trucking. There are some companies working on this problem from a few different angles right now. In theory it will get to the point where a customer can order a product, and it all gets handled by a machine until the moment it shows up at your warehouse. No humans needed! The only limit on how fast something can get to you becomes a) the speed limit and b) if a truck is available to grab your load immediately.

Right now when we want to ship something from our factory to our warehouse we call up a few trucking companies to get freight quotes on how much it will cost. Then we agree on a pick up day and they show up at the factory on that day to pick up the load. The factory has been pre-warned that the truck is coming that day, so they move our pallet of goods nearer to the loading dock, so it’s ready to be loaded once the truck arrives. This is usually done with a forklift. Then there’s what’s called a ‘Bill of Lading’ which must be signed and handed over - basically a receipt saying that the load was picked up and exactly what was in it. Then the truck closes the door and gets driving.

Imagine a world instead where I go to a freight marketplace of autonomous freight companies. I say when I need a truck, where, and where to. And the companies bid for that job, and the lowest price wins (probably the company who has a lower utilization rate at that moment in time). The factory gets auto-notified of the day of the shipment. Now they have a robot forklift which can pull the correct pallets down and lay them out “just in time” for loading on the truck. When the truck docks, maybe it scans a QR code or somehow exchanges information with the system the warehouse runs on to explain what it is picking up. Then it gets loaded, and it’s on its way with a digital BOL. This all could happen at 3am in theory. Why wait until daylight hours if no humans are in the mix? Maybe there’s a human in the loop watching 20 trucks or a single warehouse checking to make sure there are no issues, but the number of people needed to achieve this is vastly lower than before.

Truck Design and Fuel efficiency

Electrification of trucks is already underway with the Tesla Semi being out on roads, and more brands to follow. As infrastructure for electric trucks improves, I expect we’ll see a lot more of these in the autonomous era. Anything that drives down the cost of getting goods from A to B is going to be explored, and without drivers, some new opportunities open up.

Drafting, or “platooning”

Truckers already sometimes draft off of each other. Where they drive close to the truck in front in order to reduce drag on the vehicle and increase their fuel efficiency. This has been studied in depth, I found this paper (paywall) which tested different following distances across paired (or “platooned”) trucks. In a pair, both the front and back truck gain fuel savings, and together the best result was a 6.4% fuel efficiency improvement, at 55mph and 30ft following distance. This isn’t even a novel idea, there’s companies working on it already, including Platoon Tech in MountainView, although their website seems to be down so I suspect they may no longer be in business. I found this deck online from a different company, Peloton Technologies (acquired in 2019), which if you’re interested goes into great detail on the challenges of making this work with human drivers and existing trucks. Here’s one slide detailing how their system helps with stopping distances when close following by linking the trucks together wirelessly:

If we no longer have humans at the wheel, is it a crazy leap to think that driverless trucks could communicate with each other and organize themselves into convoys? Computer reaction times beat human ones, so you could chain together tens of these trucks and get them to within distances of each other that allows for the best possible fuel savings. Fuel savings = increased range = fewer stops = faster deliveries.

Aerodynamic shape

Let’s go a step further still on aerodynamics. We don’t need space for a human at all. Let’s make it as close to perfectly aerodynamic as we can. No wing mirrors needed.

Second Order Effects to look out for:

Trucking industry impact:

• The number of trucks on the road will increase as they become more efficient from a fuel and cost standpoint, stealing share from trains and planes, and likely increasing the overall size of the pie as shipping more stuff becomes feasible. Pretty classic Jevons Paradox here.

• Jobs will shift from the driver side to fleet maintenance and coordination at scale. It’s unlikely that one person would own an autonomous truck, more likely that a company would own a fleet of them.

• Truck rest stops will shift to accommodate the new reality. Likely adding battery swap stations or more electric chargers as the fleet electrifies. It’s likely we’d see more maintenance shops pop up at these rest stops too.

Regulation

• This whole industry is incredibly fragile. Even though human driven trucks have accidents daily, often fatal because of the speeds and tonnage involved, just one accident of a driverless truck could set back the technology by a decade. It’s important that early players play it safe with their rollouts.

• At first, autonomous trucks would likely be given access to specific routes along highways. Maybe they’d have human drivers hook up to trailers for the last mile delivery if the roads get more tricky to use.

• Eventually we might see autonomous truck specific lanes, and the addition of sensors / signal systems built into the infrastructure to help the fleets move through the world efficiently.

Consumers

• Cheaper, faster delivery of everything. Especially as warehouses catch up to the 24/7 nature of autonomous trucking, we’ll see “business day” shipping disappear as a concept. Warehouses and trucking will operate over weekends.

• Highway congestion is hard to predict. It’s possible that these systems could optimize for night driving when there’s less human traffic on the roads.

• Cleaner air if electrification continues. Trucking produces 23% of all transportation greenhouse gas emissions.

With all these changes, the highways of the future look a lot sleeker and perhaps characterless than those we’ve been used to historically. But maybe we won’t even notice, as the parallel change here is the development of autonomous driving in cars, already way ahead of trucks with Waymo taxis and Tesla Robotaxi’s on the road already in multiple cities.

The improvements autonomous trucking can bring to industry in general by making it both faster and cheaper to move physical “stuff” around the country are impossible to ignore. Love it or hate it, roads 20 years from now are going to have a lot of these on them, and our current driving landscape will be nearly unrecognizable. There are more second and third order effects that I’ve missed, and I’d love to hear from you if you have some future predictions in this area that are different to mine.

---

Thanks to

CansaFis Foote

and Cam Houser for feedback on drafts.