A Slow Merge for Truck Automation


It seems almost inevitable that automated trucks will be introduced to our roads at some point, but it will not happen in a single step and it will take time.

The attractions are substantial. In some future end-state, driverless trucks will remove the cost of the human operator from the cab and allow the equipment to operate almost continuously, eliminating restrictions on driver hours of operations, thereby delivering significant productivity gains. Another significant benefit is the expected reductions in road accidents, because most accidents are rooted in driver errors.

This end state is, however, very far into the future. Current autonomous vehicle systems retain the services of an operator in the cab to handle certain tasks, a compromise that I believe will be with us for some time.

Today’s “part-way” solutions are delivering some important economic benefits, including driver-assist features such as lane-centering steering and adaptive cruise control. But there are some notable restrictions.

During trips, these features function mainly on an exit-to-exit (E2E) basis – even with an operator in the cab. The reason that autonomous operations for trucks may be restricted to E2E movements, is that in city streets large trucks simply have to violate the edge of the road when taking certain turns. In other words, they must go over the white lines, violating the edicts that are built into the software.

In addition, the technology’s economic benefits depend on the US government changing the “driver hours of operation” rule. Currently, a truck driver can be on duty for up to 14 hours, but driving for only 11 of these hours. Classifying the E2E operations of a semi-autonomous truck as “on duty” but not “driving” may provide a productivity benefit.

Another feature of advanced autonomous trucking is a mode of operation called platooning – the ability of a group of trucks to “lock” electronically to each other, and drive in a platoon with a small gap between the trucks. Even in a platoon of a couple of trucks, the operation does save gas both for the lead truck (cutting the air turbulence behind it) and the rear truck (which does not have to “push” the air in front of it). Of course, longer platoons will have more benefits.

But platooned trucks may be limited to operation on “power lanes” of common carriers – where a single carrier operates many trucks on a given stretch of highway. Also, designing the trucks to operate in a platoon is equivalent, from an inventory point of view, to operating larger conveyances. Thus, instead of small, frequent shipments, the result may be larger, infrequent shipments with the resulting penalty of service and inventory carrying costs. As a result, platooning may be of limited benefit.

The mode can deliver significant operational gains if a standard is created whereby any truck can connect with any other truck on the road to platoon as far as they go on the same stretch of highway. However, this will require a universal (or at least national) standard of connectivity, communications, and sharing of the benefits.

There are several other speed bumps to consider, even as the movement towards autonomous vehicles – whether trucks or cars – proceeds at full speed.

For example, who is liable if an accident occurs: the carrier, shipper, or truck manufacturer? Should the software provider be held accountable? In the absence of a human operator, apportioning blame becomes more complicated.

Furthermore, autonomous cars may be “too cautious” and create massive traffic jams. In other words, to deal with the liability issue, the software may cause driverless cars to operate “like my grandmother,” for example, by waiting forever at intersections. The mixed operation of autonomous vehicles with traditional drivers may create a slew of operational problems, such as how respective drivers understand each other’s intentions at intersections.

There also are several technological hurdles to overcome, such as how autonomous systems will perform during adverse weather when fog or snow mask highway lanes and render signals indistinct.

And as is always the case with the arrival of a disruptive technology, legal frameworks and regulations lag the development of the technology. Governments have shown a lot of interest in accommodating driverless vehicles, but as we are seeing with the introduction of drones, the supporting regulatory process is slow and often unwieldy.

Driver-assist features such as lane-centering and dynamic cruise control are already on the way for large rigs. Platooning may appear in three-to-five years and E2E automatic operation (with an operator in the cab) a few years later, depending on the speed of regulations.

However, it’s unclear when we will witness fully automated, dock-to-dock operation. My guess is that it will take over a quarter of a century.