Uber: Paving the Way for Self-driving Trucks
How Uber is disrupting trucking with autonomous vehicles and what additional advancements can we expect in the future.
In the last years, there have been breakthrough advancements in the development of autonomous vehicles. What seemed like science fiction a decade ago is very close to being accomplished. Companies like Uber, Google and Tesla have begun the race to develop the first fully functional prototypes. But contrary to what many might think, transportation experts expect that the earliest applications will be in self-driving trucks, not cars. Trucking routes usually consist of long-haul drives on highways, which make them more predictable compared to cities [1]. In this particular industry, Otto (acquired by Uber in 2016) is one of the main players. In October of last year, Otto managed to transport 50,000 cans of Budweiser, making this the first commercial delivery by an autonomous truck [2].
https://www.youtube.com/watch?v=Qb0Kzb3haK8
Problems in the Trucking Industry
In the U.S. more than 70% of goods are transported by trucks, which translates into a $700 billion industry [3] [4]. Despite being the preferred way of moving goods across the country, traditional trucking has severe drawbacks. Truck and bus crashes are responsible for 4,000 deaths and 100,000 injuries a year, 90% of them caused by driver error [5]. Other important issues are truck driver shortages and turnover rate. In 2017 the American Trucking Association calculated that there was a 50,000 shortage of truck drivers and that yearly turnover rate was 90%. If the trend continued the shortage of truck drivers will hit 174,000 by 2026 [6]. Self-driving trucks, even if not fully autonomous, have the potential to drastically reduce these problems by cutting down the number of accidents and reducing drivers’ workload.
Figure 1: Projections of Truck Driver Shortages
The Road Ahead
With the truck Uber used for the Budweiser it reached a “Level 4” autonomy, which meant that the vehicle was “fully autonomous” in a certain domain (in this case the highway) [7] [8]. In the next couple of years, Uber planned to continue to develop the technology so that trucks were able to handle any situation they encountered. Other areas of opportunity where improving acceleration, deceleration and lane control. They were also working on predicting behavior for other drivers and dealing with changing weather [2]. Although reaching “Level 5” autonomy was the ultimate goal, in the short-term Uber would concentrate on improving “Level 4” performance. That meant that the self-driving trucks will still need a human driver especially in the cities, where the truck couldn’t operate on its own.
Figure 2: Autonomous Vehicle Levels
Uber estimates that it could reach “fully autonomous” trucks at a large scale within the next decade. Lior Ron, co-founder of Otto, expects that by that time driverless trucks would routinely drive throughout the U.S. highways [9]. This would have significant effects on the efficiency of the transport industry as it would reduce labor and increase fuel efficiency. Labor and fuel are the most important costs of operating a truck, representing 40% and 25% of the total cost respectively [10]. Even if trucks couldn’t reach “Level 5” autonomy and would still require a driver, labor could be better utilized and substantial cost reduction could be achieved.
Recommendations
Even though the self-driving truck industry may appear very attractive, it faces intense competition. Competitors such as Google and Tesla are making huge investments to develop competing technology. This industry is attracting billions of dollars from start-ups to the biggest trucking operators [4]. Given that “Level 5” autonomy is still far away from a technology and regulatory standpoint it is important for Uber to begin with driver assistance technology. Building a commercially viable prototype, even if not fully autonomous will allow the company to slowly enter the industry and be a step ahead of the competition. Incremental changes in this technology can have big impacts on fuel efficiency, labor cost and churn rate which would make it an attractive product.
The ability to have trucks on the road will also allow Uber to quickly improve the technology. By analyzing in real conditions when and how human drivers take control of the vehicle can help the company fine-tune their algorithms. The more data that can be collected, the more autonomous the vehicle will be as it will learn how to act in a wider range of scenarios. Uber has a competitive advantage compared to competitors in this regard. Having access to tens of thousands of drivers in hundreds of cities could help the company in this data collection stage. Even though autonomous trucks and cars are different, there are still conclusions that could be drawn using Uber drivers.
Even though there has been a lot of advancements in the industry, there are still many open questions. How can Uber and other competitors lobby for legislation that will allow self-driving trucks at the expense truck driving jobs? How many years will it take to produce the first commercially viable self-driving truck? Is there enough space in the industry for a few players or is it a “winner takes it all” market?
(795 words)
____________________
[1] Sage, A. (2016, October 25). Uber’s Otto hauls Budweiser across Colorado in self-driving truck. Retrieved from Reuters: https://www.reuters.com/article/us-uber-trucking-beer/ubers-otto-hauls-budweiser-across-colorado-in-self-driving-truck-idUSKCN12P13N
[2] Davis, A. (2016, October 25). Uber’s self-driving truck makes its first delivery: 50,000 beers. Retrieved from Wired: https://www.wired.com/2016/10/ubers-self-driving-truck-makes-first-delivery-50000-beers/
[3] Ockedahl, C. (2016, May 23). Six Stats That Reveal The State Of The U.S. Trucking Industry. Retrieved from Forbes: https://www.forbes.com/sites/trucksdotcom/2016/05/23/a-glance-at-the-state-of-the-trucking-industry/#427675583d84
[4] Dougherty, C. (2017, November 13). Self-Driving Trucks May Be Closer Than They Appear. Retrieved from New York Times: https://www.nytimes.com/2017/11/13/business/self-driving-trucks.html
[5] Freedman, D. (2017). Tractor-trailers without a human at the wheel will soon barrel onto highways near you. What will this mean for the nation’s 1.7 million truck drivers? Retrieved from MIT Technology Review: https://www.technologyreview.com/s/603493/10-breakthrough-technologies-2017-self-driving-trucks/
[6] Costello, B. (2017). Truck Driver Shortage Analysis 2017. American Trucking Association.
[7] Davis, A. (2016, August 26). Everyone Wants a Level 5 Self-Driving Car – Here’s What that Means. Retrieved from Wired: https://www.wired.com/2016/08/self-driving-car-levels-sae-nhtsa/
[8] On-Road Automated Driving Committee. (2016). Taxonomy and Definitions for Terms Related to Driving Automation Systems for On-Road Motor Vehicles. SAE International.
[9] Vanian, J. (2017, March 27). In 10 Years, Artificial Intelligence Will Transform Trucking, Says Otto Exec. Retrieved from Fortune: http://fortune.com/2017/03/27/uber-otto-artificial-intelligence-truck-driving/
[10] Murray, D., & Torrey, W. F. (2016). An Analysis of the Operational Costs of Trucking: 2016 Update. American Transportation Research Institute, Arlington.
Alejandro, I really enjoyed your article. Level 4 autonomy sounds similar to the autopilot function of an aircraft. In this regard, I wonder if we will first see human-assisted vehicles much like airplanes are operated today. As featured in your video, the driver only operates the truck outside the level 4 environment and spends the rest of the time monitoring, resting, etc. This approach could prompt gradual changes in legislation that would eventually permit level 5 autonomy. It would also provide interim revenue to support this developing technology. I’m excited to see what the future holds!
Alejandro, thanks for the great read. Although I agree that in the medium to long-term self-driving trucks can improve safety on the roads, on the short-term the number of accidents can be higher until drivers learn how to work with the technology properly and understand its limitations (e.g., does it work with extreme weather?). This is particularly relevant as sellers of self-driving cars tend to highlight what the cars can do and not what they can not do.
The point above also poses challenges for related industries. In case of an accident who will be liable? The self-driving car company or the driver? How will insurance companies price the insurance for these cars and who will pay for it? In my opinion, these are questions that need to be answered before self-driving cars gain considerable traction.
Links:
https://www.wired.com/story/self-driving-cars-insurance-ambiguity/
Thank you for the very interesting post, Alejandro!
I completely agree, self-driving trucks are a fascinating topic and will necessarily disrupt the transportation business. Actually, I think that we will see self-driving trucks very soon. For instance in the UK, government allowed convoys of semi autonomous trucks, known as “platooning” on public highways as of next year for testing purposes: https://www.engadget.com/2017/08/25/uk-platooning-trials/
In Germany DHL will start tests of self-driving last mile delivery vans next year: https://www.inc.com/will-yakowicz/dhl-self-driving-trucks-germany.html
In addition, very recently, the US Senate’s Commerce, Science, and Transportation Committee approved a bill called the “American Vision of Safer Transportation through Advancement of Revolutionary Technologies Act, or the AV Start Act” speeding up legislation for self driving vehicles: http://ipjournal.law.wfu.edu/2017/11/u-s-senate-speeds-up-self-driving-car-legislation/
Great article Alejandro! Other than the costs of transportation, self-driving cars and trucks could be a great improvement in public health. Some experts predict it can reduce death in traffic accidents (one of the leading causes of death in the US and globally, http://asirt.org/initiatives/informing-road-users/road-safety-facts/road-crash-statistics) to be reduced by 90% (https://www.theatlantic.com/technology/archive/2015/09/self-driving-cars-could-save-300000-lives-per-decade-in-america/407956/).
Nice work, Alejandro!
I find it interesting to consider the business case in a world dominated by self-driving cars trucks. Will Uber own the fleet of trucks and transform itself into an asset-heavy business? Or will it make sense for logistics providers, software companies, and automotive OEMs to join forces to best tackle this solution? I find the collaborative approach to be the most promising and to this point, we have seen large auto OEMs such strike partnerships with strategic partners such as logistics providers, artificial intelligence companies, and sensor companies to come up with a complete solution to tackle the complex issue of autonomous transport. One example can be seen with GM, which has invested in and/or acquired Lyft (logistics provider – https://www.reuters.com/article/us-gm-lyft-investment/gm-invests-500-million-in-lyft-sets-out-self-driving-car-partnership-idUSKBN0UI1A820160105), Cruise Automation (AI/software development – http://fortune.com/2016/03/11/gm-buying-self-driving-tech-startup-for-more-than-1-billion/), and Strobe (LIDAR manufacturer – https://www.theverge.com/2017/10/9/16448306/gm-cruise-strobe-lidar-acquisition), to name a few. The technology and industry should continue to evolve and it will be fun to track the developments over time to see who ultimately succeeds and how.
Interesting to think about autonomous trucks, i guess I have always selfishly figured it would be my daily commuting car that was autonomous. In retrospect, the constant image on every freeway are lines of trucks moving goods. For so many reasons, again selfishly, traffic patterns, reduce congestion, or even aid congestion if used properly, the autonomous truck revolution would be beneficial. One thing that comes to mind is our safety as a nation, can we prevent a foreign power from introducing trucks into America that might deliver weapons or other threats to many cities at once? And how secure are these vehicles to computer hacking? As the Tesla electric truck is perfected this has the potential to be a hackable vehicle that is powered by a small explosive. On balance there are many positive opportunities, but as touched on prior to this comment safety and thoughtfulness are key.
The issue with self-driving trucks and cars is not technology; it is the government. Self-driving cars are able to be developed, but is the world ready for it?
This market is definitely not winner takes all. All car and truck manufacturers will compete in this industry, and it will be exciting to see who comes out first and is able to sustain the lead.
Great article Alejandro. I definitely think that the world is shifting towards more technology-driven and that is fine; it is evolution. In my view, the truck industry will be the first industry to successfully include autonomous vehicles, given the several reasons you mention in your post, and in the next 5 years. However, this will not be the end. I imagine a world where there will not be any drivers at all for any type of services (e.g. taxi, ride share, rentals, package delivery). The issue will be how and where to shift all those lost jobs efficiently and fast. In the past the market and the industry have made this shift quite smoothly, nonetheless with today’s rapide improvements in technology, that shift will be tricky and more complex.