Autonomous trucks are hitting the road soon enough…
Updated: Sep 2, 2021
A brief recap
Although you are surely familiar with the generic term “autonomous vehicle”, the first thing that comes to your mind will probably be a driverless car. However, you will be surprised to know that driverless trucks may be joining the roads much sooner than their smaller sisters.
But let’s start from the very beginning.
Experiments on self-driving vehicles started in the 1920s, when a radio equipment company drove remotely the “American Wonder” – a 1926 Chandler – through the streets of New York. Since then, the research has continued at a very good pace.
Self-driving trucks entered the space in the 1950s, when platooning - various trucks travelling in line, with only the “lead” truck operated by a human driver - was being studied for military purposes. Through the years, the transport industry has increasingly discovered that autonomous trucks may have infinite applications: a 2013 study by Caterpillar Inc. and Carnegie Mellon University underlined their potential to assure workers’ safety in mines and construction sites, while nowadays the industry is mainly focused on goods delivery, where self-driving trucks are able to safely drive on highways and in the city traffic.
There are significant differences in the technology that an autonomous vehicle may require, depending on it being remotely controlled, supervised or fully autonomous. But there exists one fundamental communality that allows the magic in all cases: sensors.
In fact, three sensors are used to control the vehicle and react to the field conditions:
Cameras are placed on every side of the vehicle. In combination with advanced computer vision technologies, they are the most accurate mean of representing the surrounding space.
Radars, radio waves that hit an object and return information about its position and speed, are particularly crucial when the visibility is low (e. g. at night) and for the detection of self-driving cars.
Lidars, laser waves that use light properties to detect the position and the shape of an object, are also used to drive in low-light conditions: they manage to create a 3D representation of the environment, enhancing safety and improving the other sensory information.
The collected data are then combined to constantly assess traffic and external conditions, ensuring a safe drive. Technology-wise, the difference between remotely driving a truck or a car is not trivial. For about 90% of the time, long-haul trucks drive on highways, which makes them easier to control and avoids accidents due to distraction. The trickier part remains the urban 10%, as it is more difficult for trucks to maneuver in the city traffic. On top of this technical challenge, because of their mass and size, accidents involving trucks are potentially more fatal than accidents including cars only.
Who's active in the space?
But who are the brave and futuristic firms pursuing this great objective? Nowadays, Southern USA is the most effervescent context, where various startups and companies are testing their prototypes on the large highways across the desert. But, as stated above, not all autonomous trucks are alike, and these companies hold very differently specialised technologies. Let’s see who, what, and how:
Daimler, world's major manufacturer of commercial vehicles, has been producing partially autonomous trucks (SAE level 2) since 2015, with humans at the wheel still required to ensure efficiency and safety. High automation trucks (SAE level 4) are also being developed: in these, the system can perform the entire dynamic driving task, but only in suitable scenarios. Hence, the prompt intervention of a human driver may still be needed.
Waymo, founded in 2009 as the Google Self-Driving Car Project, has achieved good results in the robotaxi sector, and is now also testing a fleet of autonomous trucks in California, Arizona, New Mexico and Texas. Waymo has recently started collaborating with Daimler for the development of fully autonomous trucks (SAE level 5), where human presence is not needed.
Elon Musk’s Tesla is developing the “Tesla Semi”, an all-electric battery-powered Class 8 semi-truck, equipped with “Enhanced Autopilot”, which offers semi-autonomous capability. Musk thinks that the Semi will finally bring platooning on US roads, with its production already starting in 2021.
Embark Trucks is a San Francisco-based startup which, just like Waymo, is trying to apply its thorough robotaxi technology to commercial vehicles. Although their fleet of demi-autonomous trucks is already delivering goods between suburban Los Angeles and Arizona, Embark is looking forward to reaching a Level 4 automation, so to eliminate human presence almost totally.
TuSimple was founded only six years ago, but is already partnering with major firms like McLane, UPS and Navistar. This San Diego-based startup is considered the forerunner in Level 4 autonomy: TuSimple’s highly autonomous trucks will hit US roads this year, and are expected to be the first commercially available by 2024.
Alas, as any mighty challenge, some runners already had to give up on the objective: Uber’s well-known deadly accident in Arizona has obliged the company to stop their tests, while Starsky Robotics has abandoned any self-driving truck plans after a failed round of funding.
Worth it or not?
The billion-dollar question here is whether such an amount of technology development, investments, startup-generating efforts is justified by the advantages that self-driving trucks will bring to our society.
Truly, this new way of transportation can be beneficial in many ways. Firstly, it would help fighting the shortage of drivers that European and American companies are facing today (in Italy only, a shortage of around 15,000 drivers was registered in 2020): as the job market is offering more qualified jobs, the demand for trucker positions is falling. Secondly, self-driving trucks promise to bring less accidents and road deaths, as tiredness and sleepiness would not represent a threat anymore. Lastly, such a technology could save billions in transportation costs, bringing a benefit not only in the logistic sector, but across all industries and eventually to final consumers.
On the other hand, self-driving trucks could do much more than just reducing the shortage of drivers, making them the next “switchboard operators” or “film projectionists”. More in general, this problem could shake the whole traditional trucking company ecosystem. The transition to a driverless fleet would require great investments and a fast adaptation, a very hard target for severely margin-constrained firms that are fighting to survive in an ever-evolving landscape.
As mentioned already, driverless trucks would be the response to urgent problems, like the foreseen shortage of truckers, or the strengthened demand for automated freight delivery after the 2020 pandemic. However, even if driverless trucks seem a close reality like never before, such technology might still require a good amount of time before its massive spread. Developing a SAE level 4 truck will take at least four years, while level 5 is still at early stages of design.
Apart from technology’s complexity, some socio-political barriers are curbing the rise of autonomous trucks, like the public’s general fear of “unsupervised” trucks and the lengthy adaptation of laws and regulations. On the latter, Elon Mask is currently pressing US institutions to legalise platooning, the practice on which he is betting with the Tesla Semi; at the same time, TuSimple is trying to involve regulators in its development program, making them familiar with the technology and the criticalities, so to accelerate the legislative paradigm.
More realistically, we can expect that in the close future autonomous trucks will be used in a supervised mode, allowing drivers to become more productive and make journeys safer, just as airplanes’ autopilot. Another phenomenon we can easily forecast is the spread of industry’s efforts to the old continent: TuSimple will soon join European streets, running a 100-mile-long route in Sweden, in collaboration with Scania.