The autonomous trucking revolution is about to start

[funkervogt]

A large trucking company called "Embark" and a real estate investment company called "Alterra" have partnered to create an infrastructure network for autonomous trucks in the southern U.S. The core of the plan is buying large parking lots and warehouses near highways and just outside of cities where the autonomous trucks could park themselves. The plan is for the trucks to drive themselves over long, simple interstate routes, and to then park at the nodes, where their cargoes will be transferred to human-driven trucks for last-mile delivery over shorter, but more complex routes through suburbs and cities. Embark will supply and run the autonomous trucks, while Alterra will buy and run the nodes. They plan to open the network in 2024.

https://www.wsj.com/articles/self-drivi ... 1646053200

This development is in line with my predictions. The autonomous vehicle revolution will affect the trucking sector first. Machines will take over long, simple shipping routes that only go over highways. Human truckers will be responsible for "last mile" deliveries. The number of trucking jobs will drop, but the remaining jobs will be of higher quality since the humans will be able to work close to where they live (no more cross-country trips).

This arrangement could persist for decades until last mile delivery is also automated.

The automation of long-haul trucking will devastate many small towns across the U.S. and Canadian heartland, which are heavily based on servicing truckers and can almost be thought of as glorified truck stops. Once there are robot trucks that can drive all day and night, they won't need to make stops in remote towns by the interstate, which means the fuel stations will go bankrupt, along with the diners, strip clubs, convenience stores, and motels clustered nearby. That creates an economic death spiral that basically kills the town.

There will be no relief for the residents of Trump Country. Meanwhile, more of the economy and population will continue to be concentrated in and around cities.

[funkervogt]

Also, here's the press release from Embark and Alterra.

https://www.globenewswire.com/news-rele ... e-Map.html

[R8Z]

Very very interesting and it makes total sense to just automate the longest and most soul crushing part of trucking jobs. It also being the easiest part to automate makes it an obvious choice for starting up. This going to work very well.

[Yuli Ban]

I'd say it's about time, but I think the better way to put this is that we've had plenty of time to become aware of this. Whatever happens next is going to be a rough transition to a vastly different way of life, but provided the right pressures are there, it ought to work out for the best.
I'm not saying we're going to see immediate overnight effects. Rather, this will be among the first wave of automation that will be visible to everyone. We'll be able to study the effects of trucker automation and the wider response and plan how to go forward from there in other areas.

[caltrek]

A big test will come (if it has not already occurred) when the first accident fatality occurs in which an autonomous truck is at least partially to blame. A lot of people won't listen to explanations about how such trucks are actually statistically safer than human drivers. Human truck drivers and their allies may very well be tempted to say "see, we told you so." Regulators, overseen by politicians who might not always be rational, may be pressured to ban such trucks outright.

[Nanotechandmorefuture]

I guess the truckers were right to protest. Here's to seeing if there is truth to the whole reskilling people's jobs as automation takes it from them.

[funkervogt]

I just found another study about the impact of autonomous trucks.
https://www.nature.com/articles/s41599-022-01103-w

Noteworthy conclusions:

1) The first stage of the trucking industry's automation will probably take the form of the “transfer-hub” model. "This would involve an automated truck (AT) completing the highway leg of the route and human drivers undertaking the more complex suburban-urban segments at both the starting and end points of the journey. Truck ports near highways would be used to switch out the trailer from the prime mover and enable this switch at both ends."

2) There's no guarantee that the number of jobs created to satisfy the suburban-urban segments of transit routes will equal the number of jobs destroyed that involved long-haul transit routes. The overall number of trucking jobs will probably decline.

3) While the short-distance urban-suburban trucking jobs will be better in the sense that drivers will stay in small geographic areas, allowing them to go home each night, the jobs will probably pay less than long-haul routes.

"A larger volume of shorter trips may not compensate for the loss of work associated with automated long haul trips. A transfer-hub model will require drivers to shift to short haul jobs. The most common pay structure currently for “truck-load" drivers, who haul full truckloads worth of generic containerized freight for a trucking company (and not for a shipper like Walmart or Target), is payment per mile of haulage. They are often not paid for the waiting and paperwork that occurs at the beginning and end of each trip. Truckload drivers seek to maximize the time they keep the cargo moving. A shift to shorter trips would increase the ratio of stationary (unpaid) to driving (paid) time, reducing their wages per hour worked (Viscelli, 2016)."

4) The transfer-hub model will change the places where trucking jobs are available, and cause a demographic shift to the composition of the workforce. Rural towns whose economies depend on truck stops will get less business and wither away as autonomous truck refueling stations become concentrated at large "truck ports" (like those Embark and Alterra are trying to build) near U.S. cities. Long-haul truck drivers, who are predominantly white and live in rural areas, will lose their jobs, while the new trucking jobs will mostly go to nonwhite people living in or near urban areas, close to the truck ports. This transfer of jobs and wealth will worsen the cultural, political, and economic divide between the American "heartland" and its urban and coastal areas, and play into many existing stereotypes.

These truck stops currently employ about 70,000 people (See SI Table S2). On the other hand, the creation of truck ports to facilitate the transfer-hub model of AT will likely create new jobs. These will possibly involve new tasks such as switching trailers between the human and automated prime movers, offering services to human operators that were previously offered at truck stops, and maintenance and safety checks of sensors and other equipment on board automated trucks before their deployment on a route. To a first approximation, our analysis indicates that it is possible that the labor-hours lost at truck stops and other locations on highways could be compensated by new employment opportunities at transfer-hub ports (see SI Section 4 for more details). However, it is unclear whether operators in existing jobs at truck stops will be interested in or qualified for the new jobs that may arise from deployment of AT.

5) "Partial automation" is infeasible. This is a proposed model where humans keep their existing trucking jobs, but use "autopilot" systems for the monotonous, long-haul middle segments of their routes. It fails because it is the worst of all worlds: Trucking companies have to pay for more expensive trucks that are capable of fully autonomous driving, but they also have to keep paying the human truck drivers' salaries. While, in theory, a professional human truck driver and an autonomous driving system could drive safer than either alone (presumably, the human would sense the machine's mistakes and correct them, and vice versa), in reality, the human drivers would become completely inattentive and would rely 100% on the machines to drive. It would be costlier but no safer than using a fully autonomous truck with no human inside.

[caltrek]

Rural towns whose economies depend on truck stops will get less business and wither away as autonomous truck refueling stations become concentrated at large "truck ports" (like those Embark and Alterra are trying to build) near U.S. cities.

I think this is true if no other considerations come into play. All of this reminds me of Highway 5, here in California. Truck traffic has become heavier and heavier, while automotive traffic has been held at a relatively steady level. At least if my impressions over time are correct. Truck stops sometimes served as an initial economic base for communities along this route. Still, their very proximity to Highway 5 made them good candidates for further growth, especially if there were other activities in the area such as farming, recreational facilities, etc.

Also, trucks may still need to refuel. So, while diners might be impacted, refueling stations may still be needed. Electrification of vehicles will be a complicating factor. A lot will depend upon the ultimate range of such trucks. Also, if the main hubs are close enough together, then the need for recharging stations along the way will diminish.

[caltrek]

"Partial automation" is infeasible. This is a proposed model where humans keep their existing trucking jobs, but use "autopilot" systems for the monotonous, long-haul middle segments of their routes. It fails because it is the worst of all worlds: Trucking companies have to pay for more expensive trucks that are capable of fully autonomous driving, but they also have to keep paying the human truck drivers' salaries. While, in theory, a professional human truck driver and an autonomous driving system could drive safer than either alone (presumably, the human would sense the machine's mistakes and correct them, and vice versa), in reality, the human drivers would become completely inattentive and would rely 100% on the machines to drive. It would be costlier but no safer than using a fully autonomous truck with no human inside.

One possibility is that drivers would be freed up to multi-task. Inventory control could at least partially take place on the road. Such work on mobile computers could then be easily downloaded at the hubs. Drivers might also be able to spend more time planning their routes. Their hands and attention could be freed to do this sort of work. Also, one function of drivers is to periodically inspect their vehicles for damage and maintenance problems. So, problems encountered on the road, such as collisions with road debris, could receive immediate attention and evaluation.

[funkervogt]

"Partial automation" is infeasible. This is a proposed model where humans keep their existing trucking jobs, but use "autopilot" systems for the monotonous, long-haul middle segments of their routes. It fails because it is the worst of all worlds: Trucking companies have to pay for more expensive trucks that are capable of fully autonomous driving, but they also have to keep paying the human truck drivers' salaries. While, in theory, a professional human truck driver and an autonomous driving system could drive safer than either alone (presumably, the human would sense the machine's mistakes and correct them, and vice versa), in reality, the human drivers would become completely inattentive and would rely 100% on the machines to drive. It would be costlier but no safer than using a fully autonomous truck with no human inside.

One possibility is that drivers would be freed up to multi-task. Inventory control could at least partially take place on the road. Such work on mobile computers could then be easily downloaded at the hubs. Drivers might also be able to spend more time planning their routes. Their hands and attention could be freed to do this sort of work. Also, one function of drivers is to periodically inspect their vehicles for damage and maintenance problems. So, problems encountered on the road, such as collisions with road debris, could receive immediate attention and evaluation.

Why wouldn't route planning be automated?

To your last point, it would still be inefficient since the human driver wouldn't be doing anything 99.9% of the time. It would be better for autonomous vehicles to pull over whenever they had problems and to summon the nearest human repair crew.