Consider high-profile news reports about autonomous vehicles from the last 12 months: an Uber self-driving car killed a pedestrian in Arizona, General Motor’s (GM’s) self-driving cars were involved in six crashes in California in September, and GM was sued by a motorcyclist after a collision with an autonomous-driving Chevrolet Bolt. Reading these stories, you may feel pessimistic about the future of autonomous vehicles.
However, pessimism might be misplaced, similar to the negativity of motorized vehicle naysayers during the early 1900s. These naysayers focused on the material issues of early automobiles, which were unreliable and inconvenient—they regularly got stuck in the mud, experienced multiple flat tires a day, and required someone to walk in front of them with a red flag in certain cities and towns. Many people thought that automobiles would never be more than a novelty. But, with 97 million vehicles produced in 2017, they were clearly wrong!
The Society of Autonomous Engineers defines five levels of automated vehicle evolution: level 1 (i.e., driving assistance) provides some basic features, level 2 (i.e., partial automation) provides acceleration and steering support, level 3 (i.e., conditional automation) requires drivers to take control if needed, level 4 (i.e., high automation) provides automation of all driving functions under certain conditions, and level 5 (i.e., full automation) presents vehicles that can autonomously perform all driving functions under all conditions. Though every autonomous vehicle company is striving to achieve level 5 automation, American consumers are only able to purchase level 2 (“partial automation”) vehicles, while European consumers are able to purchase only one level 3 (“conditional automation”) vehicle, namely Audi’s Traffic Jam Pilot equipped A8, which is level 3 only up to 37 mph.
As automotive manufacturers progressively—and inevitably—innovate toward level 5 automation, the transport and transportation ecosystem is profoundly changing. In this blogpost, we look at four key aspects of this change: ride-sharing fleets, mass transit, long-haul trucking, and city planning.
Levels 4 and 5 automation not only eliminate the need for drivers but also reduce the need for consumers to purchase personal vehicles. Instead of owning a car, consumers can use on-demand transportation provided by a group of ride-share companies comprised of current ride-share players and existing carmakers. Why, you may ask, would ride-sharing fleets be the outcome? First, a typical car in the United States (US) is parked about 95% of the time, and consumers look for parking spaces 0.5% of the time. So, automation enables higher load factors for automobiles. Second, the elimination of drivers means that operating margins could increase to 20%, which is more than twice what carmakers generate right now. Third, level 5 automation requires significant investment in research and development, as well as field testing and approval. So, an oligopoly of ride-sharing fleets would be the best economic structure to support vehicle automation because it provides the lowest cost per mile to consumers.
In fact, this is already starting to happen. Waymo recently launched its Waymo One Service in Phoenix, where Waymo drivers supervise self-driving Chrysler vans, GM is launching a self-driving car fleet in FY19 with a $2 billion infusion of capital from SoftBank, and Ford plans to launch a fleet of thousands of self-driving cars in 2021. Moreover, Goldman Sachs predicts that robo-taxis will help the ride-hailing and ride-sharing business grow to $285 billion by 2030.
Just like personal vehicles, the automation of mass-transit buses will bring obvious benefits, such as higher reliability, increased safety, and lower costs. However, the improvement in utilization and congestion will come once on-demand buses embrace dynamic pickup and route schedules. The routes of these buses will be based on the real-time needs of riders and the mass-transit system, unlike the fixed transit routes of most mass-transit systems today. Automated level 5 buses are already being tested in several cities, albeit at low speeds. Stockholm is running a six-month pilot test of EasyMile 12-person shuttles on a 1.5 km route in the city center. Similarly, Japan is introducing driverless buses to shopping centers, airports, and university campuses with the goal of using a fleet to connect mass transit to 2020 Olympic venues, and Shenzhen, China has begun testing self-driving electric minibuses on a 1.2 km loop in the city’s Futian District.
Level 4 automation of long-haul trucking is likely to occur even sooner. Compared to passenger cars and mass transit, trucks are driven mostly on open roads and freeways, and even though there are more than three million truck drivers in the US, long-distance freight trucking costs have increased sharply due to a shortage of drivers and increased safety monitoring. Leading truck and technology players have taken note of the business opportunity here. Volvo uses autonomous trucks to haul stone out of a mine in Norway on a designated route. Daimler, after prototyping automated trucks on public roads in Nevada for three years, has announced a $600 million investment in automated driving features and electric drivetrains. Meanwhile, Uber is working on a self-driving kit for commercial trucks that will allow drivers to sleep during long-haul trips.
The advent of ride-sharing fleets and automated mass transit enables the reimagining of city planning, which will hopefully reverse the trend that has appeared over the last century: increased vehicle congestion, less use of public transport, and increased urban sprawl. Rapid Flow technology, spun-up from Carnegie Mellon University’s intelligent transportation systems, “optimizes traffic flows every second by adapting to real-time change in traffic.” In test conditions, this has led to a 25% reduction in travel times, a 40% reduction in the time spent waiting at signals, and a 30% reduction in emissions. Similarly, the Karlsruhe Institute for Technology has conducted simulations that show a lasting 30% reduction in travel times in fully automated traffic.
Where does regulation fit into the automation journey? History tells us that the commercialization of any innovation, especially if human life is at stake, always comes with material regulation and certification requirements. For example, the Boeing 787 Dreamliner went through 20 months of flight certification and 200,000 hours of FDA technical expert review before deployment; likewise, medical device manufacturers go through a rigorous, multi-year certification process with the FDA before being certified for sale.
To date, self-driving efforts on public streets have utilized vehicles that are Federal Motor Vehicle Safety Standards (FMVSS) certified with no material certification or validation of the self-driving technology. As level 3–5 vehicles become a commercial reality, FMVSS are likely to be amended, with stricter standards for safety, testing, and certification.
Having said that, common use of levels 4 and 5 automated vehicles is a matter of when, not if, although this is not likely to start in your garage. As Bill Gates stated in his 1996 book, The Road Ahead, “We always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next ten.”
We would like to thank Jarett Smith, VP of Digital Product @Disney and a guest lecturer at USC Marshall School of Business, for generously donating his time and insights to this blogpost, and we would also like to thank Atif Rafiq, CDO and Global Chief Information Officer for Volvo Cars, for his comments.