Autonomous truck

A futuristic example of an autonomous, cabless truck capable of multiple load types.

Autonomous truck refers to using artificial intelligence and robotics to develop autonomous long-driving trucks to haul trailers and freight faster, farther, and safer than human drivers can achieve. This is largely due to the restrictions placed on human drivers on how long a human driver can drive continuously and how long they have to take a break after that period. Autonomous trucks are expected to drive continuously regardless of time of day and without succumbing to exhaustion. For example, one test of autonomous trucks found an autonomous truck was able to make four round trips and deliver eight loads of freight in a five-day trip, which would have taken a human driver more than ten days to deliver the same freight. Other than productivity, autonomous trucks are expected to bring improvements in fuel efficiency, costs, and traffic on the highways.

Further, there is an expectation that autonomous trucks will be in use before other autonomous vehicles because most of the driving semi-trucks do is on highways, which present highly regulated environments with fewer driving variables than other driving environments, such as in cities.

Autonomous trucks are designed to offer improvements in fuel economy over driven trucks due to the need to speed on highways or overtake other trucks. Often human drivers will feel the need to speed or overtake other vehicles to meet their deadlines for shipping goods, with the added stress of the limitations on their drive time. Meanwhile, an autonomous truck, able to drive without restriction, does not need to worry about driving over a posted speed limit, potentially leading to fuel economy savings of around 10 percent or more, depending on the calculation. The reduction in variation of speed and the expected reduction in side-to-side movement in driving is anticipated to reduce tire wear and tear on the truck overall, increasing the longevity of the platform while reducing necessary maintenance and repairs.

Example of early autonomous trucks which offer human drivers an autopilot feature.

The actual equipment is projected to cost more than a traditional Class 8 tractor. However, the expectation is these robotic or autonomous trucks would pay for themselves over time, based on increased efficiencies and the ability to run the trucks for longer, without concern for downtime required with human drivers.

Autonomous trucks will have to overcome several challenges for the technology to become a mainstay on the roads. These challenges, or barriers, will depend on the region in which the autonomous trucks will operate, but they tend to include regulations, road conditions, merging on and off the highway, and the related infrastructure. One challenge that will have to be solved will be cybersecurity of the trucks, as there could be a risk of these trucks being hacked.

Arguably, one of the biggest hurdles facing autonomous trucks will be the regulatory environment. In the United States, at least half of the states allow autonomous trucks, but the other half either explicitly do not allow autonomous trucks or are not clear in their regulations around them. California, for example, is seen as a desirable state to allow autonomous trucks to drive through, given the state's large ports, and it has seen a bill tabled that would limit or disallow autonomous trucks. Although, the authors of the bill have noted that it is not meant as a bill to ban autonomous trucks forever, but only until the vehicles and technology have been proven safe and capable.

There is an expectation that regulators will ease their concerns as the technology continues to prove vehicle safety, capability, and the ability to drive short- and medium-length routes that are hyper-constrained, structured, point-to-point, fixed, known, and repeatable. Especially as autonomous vehicle manufacturers prove the highly repeatable and controllable nature of those routes, ease the safety concerns, and have human drivers taking over in difficult conditions. One such use case has been compared with aircraft pilots, in that a truck driver could get the truck on a highway before putting the truck on autopilot and only taking over in the case of an emergency, inclement weather, or at the end of the trip.

Weather conditions, unforeseen events, and general conditions of the roads on which the truck drives can make a difference in the effectiveness of autonomous vehicles. These conditions are a part of driving, and automated trucks have been shown to have difficulty with inclement weather, especially in the case of snow and heavy fog when visibility is limited and the ability of the technology to find details can be challenged.

Similarly, there are questions about how autonomous vehicles will deal with unforeseen events, such as sudden accidents or reckless drivers, which offer variables an autonomous truck may not have encountered or may be unable to navigate without damage to itself or vehicles around it. However, as developers of the technology have pointed out, more time on the roads dealing with human drivers, erratic driving behaviors, and other unforeseen events will allow these systems to better predict and deal with these events.

Another concern can be the condition of the very road itself. Potholes, oil slicks, poorly lined roads, and different road surfaces can all change how a vehicle handles and how an autonomous vehicle reads the road. Potholes, in particular, can cause unforeseen events, including popping tires or causing other damage to a vehicle that an autonomous system may be unable or untrained to understand potential damage. This can especially be a concern if an autonomous truck is pulling a traditional trailer without integrated sensor systems.

Another technical challenge to full autonomy can be how these trucks will merge on and off the highway and how the autonomous truck will operate in points of origin, such as at a warehouse or port; how they will operate during refueling or in the case of roadside maintenance and repair; and how they will operate at the point of destination. Part of this could be solved by having a human driver onboard.

DB Schenker has tested driverless yard logistics for autonomous trucks with teleoperation.

As trucks move toward full autonomy, there is an anticipated overhaul to the logistics infrastructure that will be necessary. For example, ports can be overhauled to have specific loading areas for autonomous trucks, where the technology will be capable of communicating with the truck and directing the truck as necessary. Similarly, warehouses and logistics centers are going to need to change their infrastructure to accept and interact with autonomous trucks. And in the case of deliveries to specific buildings, not warehouses, these trips will be more challenging for autonomous trucks to complete. However, some have noted that solving these challenges can lead to greater efficiencies in these systems, many of which have not been fundamentally improved since they were developed in the early twentieth century.

A truck with a sensor array to allow it to "see" the road.

Similar to other autonomous technologies, autonomous trucks rely on an array of technology. This includes sensors—such as cameras, LiDARs, and radars —which feed data to a computer and artificial intelligence engine, which is trained on massive amounts of driving simulations and controls the vehicle through the trip. Trucks offer a better platform for autonomy in vehicles because their large size provides more power for computers and a better field of vision for sensors, as trucks sit higher off the ground than other vehicles.

The improved field of vision is important, as trucks need to stop sooner as they require longer to stop. These sensor systems can use a combination of technologies to locate and recognize other vehicles and calculate their trajectories further out than human drivers are able to see.

An important component in autonomous driving, as noted above, is the use of artificial intelligence, which aims to create systems that can react and interact with a variety of situations, particularly unexpected situations. The aim is to develop AI systems that can react in similar ways to humans, especially as others on the road will often be human drivers, and also react faster and with greater consistency.

Another technology that will help autonomous trucks to operate is Vehicle to Everything (V2X) communication, which offers hyper-connectivity with anything connected around them. This can include other vehicles, traffic lights, GPS location data, cloud-based information resources, government hazard warnings, traffic conditions, and any other devices that can be included in an Internet of Things (IoT) environment. The ability to use a variety of connected mobility approaches can help autonomous vehicles sense their surroundings and could be used in the case of a truck requiring roadside assistance or included in an accident.

The use of telematics will allow operators to connect their fleets and orchestrate multiple trucks through shipping operations. It will also offer operators real-time location and route tracking, information on vehicle speed and performance, data on vehicle behavior over time, and diagnostic signals from sensors and ECUs. This could allow operators to collect the necessary data for fleet maintenance and management and ensure that operators can keep track of assets in the case of accidents or other unexpected events.

Despite the above benefits and the research into autonomous truck technology, concerns remain around the safety of the technology. Especially as trucks are large vehicles and cannot maneuver around a potential accident the way a car can, with a truck requiring more room and time to come to a complete stop when braking.

There are also potential problems with the sensors on a truck's cab, as the position of the sensor could find itself blinded by the sun, have difficulties distinguishing between cars and large signs, and become impaired by inclement weather. Further, there are concerns with the potential performance of autonomous trucks in city locations where there are constant stops, turns, and tight spaces and issues related to autonomous trucks merging on and off highways.

There are also concerns about the interaction between autonomous trucks and humans. This can include the flood of other vehicles surrounding an autonomous truck with various levels of autonomy of their own and fully-human driven vehicles. Anticipating the actions of these human-driven cars can be difficult for an autonomous truck, as people can be hard to predict for an autonomous system. And in the case where an autonomous truck can communicate with other vehicles around it, the human-driven vehicle could present a further risk.

Some are concerned about the potential for damage to other people's property, especially as trucks carry a lot of energy at high speed and this could cause damage to other people's property, and, in turn, risk damage to a person. Often the damage is considered to property on the road, but in the case of a truck driving on country highways or city streets, this could extend to other types of property an autonomous truck could damage.

In consideration of property damage, is the risk autonomous trucks run in terms of the loss, damage, or theft of the truck or its contents. Vehicles will need to stop for refueling at given points, and without a driver present, they present an opportunity for theft. Similarly, in the case of an accident, there is the potential for the truck and its cargo to be damaged. Also is the issue of how an autonomous truck will signal its operator that it has been in an accident.

Another concern is what happens if an operator loses control or track of the autonomous vehicle, especially in the case that an autonomous truck deviates from a planned route or does something else unexpected that makes tracking the autonomous truck down difficult or impossible. This could be a result of a failure in the vehicle's programming, a coding or data error by the route programmer, or even a power failure that results in a reset of the system and loss of necessary data for the autonomous truck to arrive at its final destination.

In the case of autonomous trucks, and before autonomous trucks reach full autonomy, there will be a need to develop specific routes for autonomous trucks to ensure they can operate at full capacity. Some of the concerns in routing autonomous trucks include the legislative environment (for example, driving through a state with legislation outlawing autonomous vehicles would not be ideal), lengthy limited-access highway shipping corridors, the freight volume, and the weather.

Much of the testing for autonomous trucks in the United States has occurred in the corridor between Dallas, Austin, Oklahoma City, Atlanta, and Miami. There are specific highways that offer smooth driving, helping the trucks operate at their best while offering favorable distances and legislation for the trucks to drive through. Favorable highways can include a higher density of truck stops and warehouses to reduce the potential "dead zones" where a truck could be lost to tracking technology and to ensure the truck can be refueled.

There are several anticipated phases of autonomy in the development and deployment of autonomous trucks. Each phase is differentiated based on the level of autonomy the vehicle has and whether there needs to be a driver in the truck's cab.

Phase one is expected to use a technique called platooning, in which a fleet of trucks will follow a lead truck on the highway. The early platooning is expected to include drivers in all the trucks and would allow the companies developing the autonomous technology to have humans onboard to locally monitor the performance of the autonomous system. With the technology expected to come in waves, this is expected as the first, in which autonomous companies can navigate challenges in the technology and regulatory landscape with proven capability. Some autonomous vehicle companies have already surpassed this early stage in their testing.

Example of a platoon of three trucks.

Phase two will be expected to have developed enough to have a driver in the lead truck of a platoon but not in the convoy of autonomous trucks following behind. These platoons of two or more trucks are anticipated to appear on highways between 2025 and 2027, depending on the estimates. And they are expected to relieve a shortage of qualified long-haul truck drivers, due to the lack of interest in younger generations to drive long-haul trucks and the increasing retirement of qualified drivers.

Phase three continues the platooning of trucks, but in this case, the lead truck is completely autonomous on the highway, albeit with a human in the cab for the navigation of small roads and loading docks. Or for safety concerns. This phase, also called constrained autonomy, would see autonomous trucks operate throughout interstate-highway systems or other "geofenced" areas without a platoon, with expectations that the vehicle's ability to operate would be subject to weather and visibility conditions and infrastructure developments (especially in the case of V2X communication technologies). In this scenario, it is likely that human drivers would meet the truck at an interstate exit and drive them to the ultimate destination, including through navigating city streets, parking lots, and loading docks. This is anticipated to occur in the early 2030s at the most optimistic guess.

Example of a driverless truck integrated with existing trailer infrastructure.

Phase four is defined by completely driverless autonomous trucks on the road and at scale. The final phase is projected to take the longest to develop and implement and will be dependent on the performance of the previous waves (as, if there are a lot of accidents or other roadblocks in the development, this phase may take longer to implement). In this phase, trucks could be delivered without a cab and realize the full potential of autonomy, if not more. This would also be a deployment of Level 5 autonomy in vehicles. This would also take full infrastructure changes.

One potential step in the autonomous trucking development has been the suggested use of teleoperators. This could be used in the case of surprise inclement weather or other road conditions that are unexpected by the operator. Similarly, teleoperators have been suggested as a potential solution for bringing autonomous trucks from a highway, through a city, or to a warehouse or delivery location without needing the truck to hook up with a human driver at a waypoint—as has been suggested in the past. This would allow the potential take-over of an autonomous, cabless truck by a human operator when the conditions require it.

Teleoperator training with an autonomous truck.

For this technology to work, data delivered over mobile networks would have to increase, with some suggesting full deployment of 5G technologies would allow for it; or, if 5G is shown to not be able to operate the teleoperator at the speed necessary, the development of 6G networks and their proposed networking capabilities would better allow for teleoperators. This is particularly as any potential lag between the operator and the vehicle—introduced over the network—could have disastrous results to human life and property.