The Road to Automation

ITF’s Landmark Report Highlights The Risk of the Digital Age on Fleets

Posted in Industry Updates

The Managing the transition to driverless road freight transport report produced by the International Truck Forum explores how a transition to driverless trucks could happen. Reduced reliance on humans to move road freight offers many benefits. It also threatens to disrupt the careers and lives of millions of professional truck drivers. Based on different scenarios for the large-scale introduction of automated road freight transport, this study makes recommendations to help governments manage potential disruption and ensure a just transition for affected drivers.

Three leading transport-sector organisations joined the International Transport Forum for this project to assess benefits, costs and risks of a transition to driverless trucks. The International Road Transport Union, the International Transport Workers’ Federation and the European Automobile Manufacturers Association contributed insights on driverless technology in the road freight sector as well as funds for the research.

Driverless trucks could be a regular presence on many roads within the next ten years. Self-driving trucks already operate in controlled environments like ports or mines, and trials on public roads are under way in many regions including the US and the European Union (EU). Manufacturers are investing heavily into truck-automation technology while many governments are actively reviewing their regulations to understand what changes would be required to allow self-driving vehicles on public roads.

Automated trucks would enable cost savings, lower emissions and safer roads. They could also address the emerging shortage of professional drivers faced by the haulage industry, particularly in Europe. Without driverless trucks, around 6.4 million truck drivers are projected to be needed across Europe and the US by 2030, yet fewer than 5.6 million are projected to be available and willing to work under current conditions. The majority of truckers are in the later stages of their careers, while few women and young men are choosing trucking as a profession.

The adoption of driverless trucks is likely to reduce demand for drivers at a faster rate than a supply shortage would emerge. Of the 6.4 million driver jobs in 2030, between 3.4 and 4.4 million would become redundant if driverless trucks are deployed quickly. Even accounting for prospective truck drivers being progressively dissuaded by the advent of driverless technology, over 2 million drivers across the US and Europe could be directly displaced by 2030 in some of the scenarios examined for this study.

Preparing now for potential negative social impacts of job losses will mitigate the risks in case such a rapid transition occurs. While truck drivers are typically flexible, self-reliant and able to concentrate for long periods, their relatively low education level and potential automation in other sectors puts them at a high risk of extended periods of unemployment. Support available to displaced workers in developed economies may prove to be inadequate given the potential speed and scale of job losses. Active management of the transition will likely be needed to smooth the introduction of driverless technology, avoid excessive hardship for truck drivers, and ensure the gains from the technology are fairly shared across society.

Governments, industry and researchers should continue to advance tests on public roads in designated corridors and areas for trialling vehicles, network technology and communications protocols. This way various technologies are able to be tested without committing to an individual company, standard or technology early in the development process, ensuring that expensive network-wide investments are not wasted or over-specified. This will help ensure societal benefits from automated road freight transport will be maximised.

Harmonisation of rules across countries is critical for maximising the gains from driverless truck technology. Common vehicle standards and operational rules would allow smooth cross-border movements of autonomous trucks and should be put in place at least at a continental level, preferably at the global level. The proactive approach of many governments to test permits and ad hoc exemptions to road rules allows different approaches to be tested in parallel which can speed up the maturing of the technology. However, such competition entails the risk of insufficient attention on the ultimate goal of harmonisation.

Governments should establish a transition advisory board for the trucking industry to advise on labour issues associated with the introduction of driverless trucks. The board should be temporary and include representatives from labour unions, road freight businesses, vehicle manufacturers and government. It would support the government in choosing the right policy mix to ensure that costs, benefits and risks from automated road haulage are fairly distributed.

Governments should consider a mechanism to shape the transition to driverless trucks. A permit system would offer influence over the speed of uptake as well as revenue to support displaced drivers. Where economy-wide unemployment support is considered inadequate, additional assistance could come in the form of targeted labour market programs to try to re-deploy drivers. It could also take the form of additional income replacement payments where alternative employment opportunities have also been reduced by automation. For reasons of fairness, funds for transition assistance should be generated by the main beneficiaries of the operation of driverless trucks. The sale of permits to operators experiencing operating cost reductions could be complemented by contributions of all road users who will benefit from improved safety. Careful design of the permit system would ensure that permits are used to manage the labour transition fairly and not as a proxy to limit the free movement of goods.

Industry transition

Adoption of driverless trucks would have profound implications for road freight operations and costs. Though this study does not to seek to provide forecasts of cost savings, some indication is useful to understand the business imperative for the development and adoption of such technology. Furthermore, this looks only at the on-going operating cost savings after transition costs (such as capital purchase and redundancy payments to former drivers).

Though some drivers operate for relatively poor remuneration, labour costs still account for 35 to 45% of the costs in the road haulage sector in Europe. Similarly, in the US, driver wages and benefits account for an estimated 35% of marginal (per mile) costs of freight operations (Torrey and Murray, 2015). Of course not all of these labour costs would be eliminated by the adoption of driverless operations, as some driving and non-driving tasks would likely remain. However, even if half of these costs could be avoided through the introduction of driverless trucks to some supply chains, operating costs could be nearly 20% lower.

Furthermore, road freight operators are facing a shortage of suitable drivers in the US and Europe, which puts upward pressure on driver wages and operating costs (as detailed in “The potential scale of truck driver job losses” section; Costello and Suarez, 2015; Samek Lodovici et al., 2009). This issue is particularly acute in the long-distance segment where long periods away from home and the potential for boredom do not offer an attractive employment proposition.

Fuel efficiency is expected to improve as braking and acceleration commands are optimised (e.g. adaptive cruise control) and improved aerodynamic performance is achieved (platooning). Estimates of the fuel savings for automated functions are in the order of 4 to 10% for automated “eco-driving” of non-platooned trucks and 6 to 10% for partly manually driven platooned trucks.

The combined effect of full automation and platooning could exceed 10%, though most of these gains are possible without driverless operations per se. In the EU, ERTICO and the European Automobile Manufacturers’ Association (ACEA) have initiated a new project that is exploring the fuel and emissions impacts of automated vehicle functions for commercial vehicles
“ITS4CV” (Intelligent Transport Systems for Commercial Vehicles).

The number of hours in a day that a human driver can operate a truck is limited by both physiological abilities and government safety regulations. Particularly in long-distance trucking, the constraints on shift lengths mean that vehicles can be sitting idle for a significant proportion of the day (unless operating as a driving team).

The introduction of driverless truck technology would remove this major constraint and potentially enable much more intensive use of the vehicle fleet (depending on other regulations, such as night-time operating curfews in urban areas). The extended hours of operation of vehicles would bring substantial cost savings as a given task could be done with a smaller fleet. The extent of the reduction in fleet would depend on the context but for long-distance tasks a reduction in the order of 50% is plausible. The actual cost savings might be relatively small since the fleet would require more frequent maintenance and replacement, and there may be some incremental costs to equip vehicles with driverless technology.

Perhaps more than 90% of road crashes in Europe and the US are due to human factors (Frisoni et al., 2016; Singh, 2015). So there is the ambition and expectation that widespread adoption of automated vehicles would reduce the number of crashes, deaths and injuries on the roads, particularly as some specific crash causes should be eliminated altogether (e.g. falling asleep or driving under the influence of drugs). However, it is very difficult to know what the actual crash performance of automated systems will be since new crash types could emerge, and other new insurance risks may involve serious costs (e.g. hacking or theft of “unattended” cargo).

In any case, if and when an improved crash (and overall insurable cost) performance of driverless vehicles can be demonstrated to insurance companies, premiums can be expected to significantly decrease (AXA, 2015).5

Taken together, ITF’s analysis suggests that a reduction in operating cost from adopting driverless trucks is possible in the order of 30% compared with today’s costs. Morgan Stanley (2013) estimates a potential savings to the (overall) US road freight industry of USD 168 billion annually. Beyond the unit cost reduction, the extension in the daily range of a freight vehicle would significantly improve the delivery times offered by long-distance road freight. This degree of cost-quality improvement explains industry’s strong interest in the technology, in spite of the many challenges that still remain and the R&D costs required to resolve them.

A wide range of technologies has been introduced into cars and trucks in recent decades. For example, anti-lock braking systems, which are now standard in new cars sold in the EU and US, have been shown to improve on-road stopping distances (NHTSA, 1999). In-vehicle navigation units have removed the need for drivers to consult paper maps. Together these and other developments have improved the safety and ease of driving and helped improve the labour productivity for the road freight industry.

More recently, technologies that can further support, and even take over some aspects of, the driving task have been made available in cars and trucks. Driver assistance systems currently deployed in new vehicles are capable of monitoring blind spots when changing lanes, automatically manoeuvring a vehicle into a parking space, and adapting the vehicle’s speed to a safe distance from the vehicle in front (Frisoni et al., 2016). At the same time, driverless truck systems have been deployed on mine sites in Western Australia and at the Port of Rotterdam (Diss, 2015; Allen, 2015). Trucks are currently also being tested on the interstate highways of Nevada in the US, where the driver is only required to take control of the vehicle in an emergency or when changing lanes (Grobart, 2015).

Research and development in the automotive industry (and broader technology industry) is currently directed at technology that can take over even more aspects of the driving task. Although significant progress is still required before fully driverless operation on the open roads could be deployed, such technology is at least a realistic prospect in coming decades and therefore demands attention.

Many studies have explored the complex technological and regulatory issues associated with a wide variety of automated vehicle technologies, especially for private cars. However, fully driverless truck technology is a specific challenge with somewhat unique motivations and impacts. In particular, driverless trucks would have a highly disruptive impact on the lives and careers of current (and future) heavy vehicle drivers. So there is a clear need for an evidence base and a plan to manage disruptions to people’s lives and livelihoods if and when driverless trucks are taken up. As trucking roles evolve, it will also be important that the industry can equip its people with the right skills.

Further, restrictions on the time a driver can drive for over a given day or week limit the speed and reach of long-distance road freight, where individual drivers are allocated to each truck. At the same time, road freight operators can struggle to attract drivers to undertake such long-distance trips.

Clearly the possibility of dramatically reducing labour input costs and relaxing the driving-time constraints on vehicle productivity would be of great interest to road freight businesses and their ultimate customers. More broadly, driverless truck technology offers the possibility for improved safety, fuel efficiency, asset utilisation and environmental performance. However, the timing and regulatory acceptance of driverless truck technology is still highly uncertain.


Trucks operating without a driver could undertake long-distance freight much more quickly without working hour limitations, substantially increasing a truck’s daily range. Together, these would have significant effects on demand, not just for road freight, but potentially for freight overall.

Road freight currently accounts for nearly half of the tonne kilometres undertaken in the EU and US. Road freight’s success in specific market segments over alternative modes like shipping and rail can be more to do with reliability, distance and network reach rather than price and speed. Nevertheless, an improvement of the scale considered here would mean that other parts of the freight market could be more strongly contested by road, unless competing modes are able to exploit the possibilities of automation to find comparable cost savings.

Significant overall reductions in the cost of freight – and an extended daily range – could lead to the adoption of more transport-intensive production models (e.g. decentralisation and increased specialisation). This increased demand for road freight could result in an expansion of demand for labour inputs in the sector (and elsewhere in the economy). Detailed modelling would be required to understand the net impacts on road freight demand (particularly in the presence of carbon dioxide emissions pricing) and employment. For example, the increase in road freight demand spurred by automation in the long-distance segment could result in greater demand for human drivers in urban areas and an increase in off-peak road freight.

A recent example of this type of feedback was in retail banking when automatic teller machines (ATMs) were introduced to undertake some of the functions of human tellers. Bessen (2015) examined employment and branch data in US retail banking and found that the operating cost reductions derived from having ATMs and fewer staff per branch encouraged banks to open more branches. Total retail banking employment stayed steady over the period in which 400 000 ATMs were rolled out.

Automation of driving tasks means that they are undertaken by computer-based systems rather than a human driver. Automation can be described either in terms of automation features, e.g. “can a system automatically regulate a safe distance to the vehicle ahead?”, or in terms of capabilities, e.g. “can a collection of systems conduct the overall driving task without human intervention?”
The AdaptIVe industry and EU research initiative takes the features approach and has developed a full conceptual framework that describes and names all automation building block features, such as parking assistance (Bartels et al., 2014). In contrast, SAE International (2014; 2016), an international association of engineers, developed a framework for describing the overall capabilities of vehicles. The widely accepted SAE framework identifies levels of automation from “no automation” (level 0) to “full automation’ (level 5) based on the extent to which the major functions of the driving task are automated, as well as the contexts and situations in which a human driver is required to take control, i.e. the system’s “operational design domain” (Figure 1).
Intermediate levels of automation (e.g. level 3) may require drivers to take control of the vehicle only very occasionally in the event of an emergency. In such a situation, it is possible that drivers would not tire as quickly as a driver that is undertaking most of the driving tasks. While it is technically possible that the length of driver shifts could be safely extended, project stakeholders were strongly of the view that the rules governing the length of shifts would not change in response to the availability and adoption of such “conditional automation” systems.

All lower levels of automation will always require a human driver to be able to take control of the truck. There would be no significant labour implications unless the technology reaches a stage where drivers are not required to be on-board the truck. SAE levels 4 and 5 describe fully automated driving. Here a vehicle’s on-board systems can collect and respond to sufficient external information to allow the vehicle to safely operate without human input. The key distinction between the two levels is that level 5 describes a set of systems that is able to automatically operate the vehicle in any situation (ITF, 2015).

Register your visit to Automechanika Dubai 2019