Autonomous Driving
| Tags | OutlookStreetTechnology📕 |
|---|---|
| Publication Date |
Content
Overview
Autonomous driving refers to vehicles which are able to autonomously navigate streets without any (or barely any) input from a human driver. The extent to which a vehicle is able to drive autonomously can be classified on a six-level scale, ranging from zero to five. In doing so, each of the six levels signals a decreasing dependence on human input, ranging from no human assistance at all (Level 0) to fully autonomous driving in any area or situation (Level 5) [1]. Until a vehicle reaches autonomy level 4, drivers are still required to actively monitor the vehicle. Hence, truly autonomous driving, as commonly understood, refers to level 4 or level 5.
Level 4 autonomy refers to full autonomy in geo-fenced areas whereas level 5 refers to full autonomy in any situation [2]. In comparison to level 5 autonomy, this means that autonomous cars only need to be familiar with a predefined area which makes navigation more predictable. While level five autonomy might take another decade to be market ready, level 4 autonomy is already employed for specific services like self-driving taxis in Phoenix, Arizona [3][4].
Whereas most cars on the streets today employ level zero or level one autonomy, modern driving assistants like Tesla's autopilot are able to reach level two [5]. As discussed above, level two systems are able to provide steering and acceleration support. They still rely on drivers’ supervision, however.
Currently, there are numerous start-ups and big corporations developing self-driving technology, most notably big tech companies as well as major car manufacturers. Waymo, which belongs to Google’s parent company Alphabet, is leading in terms of disengagements by miles driven [6]. Amazon and Apple are investing in self-driving technology as well [7][8]. Currently Apple is rumored to debut a self-driving car by 2025, a timeline that is seen as very ambitious by industry insiders [27]. As for OEMs, the most ambitious projects (as measured by autonomous miles driven) are led by General Motors as well as Volkswagen and Ford with their Cruise and Argo.AI initiatives respectively [6].
Potential for Autonomous Driving
Compared to regular cars, autonomous driving technology has multiple benefits. For instance, self-driving cars could give drivers more time for leisure or work, since they no longer have to steer the vehicle [9]. As a result, autonomous cars could become so-called third places or extensions of first places (home) or second places (work). Third places are distinct in that fall neither in the area of home nor the area of work. They are places for “informal public life”, like cafés, libraries or hair salons [10]. This could increase quality of life and make commuting less of a chore, if time spent in a car is used for work or recreation. Combined with business models like autonomous ride hailing, this technology makes commutes even more convenient. So-called robo-taxis could also make people reconsider whether car ownership makes sense for them [13], which could reduce the rate of car ownership.
Autonomous cars could also help reduce accidents and therefore improve road safety. According to Statistisches Bundesamt, in Germany 67% of road accidents with personal damage are caused by driver misconduct [11]. In the US, Eno Center of transportation estimates human error to be the primary cause of 93% of all car accidents [12]. With self-driving cars, these accidents involving errors in right of way, turning or alcohol could be minimized. To put this in numbers, it is estimated that if 90% of all cars on American roads were self-driving, the number of car accidents could be reduced by over 4 million instances per year [12].
Moreover, this reduction in accidents could help reduce congestion. Studies suggest that car accidents cause between 20% and 25% of congestion on German highways [14]. Therefore, minimizing the number of car accidents could also eliminate a significant amount of congestion, which in turn would reduce average travel time.
Finally, a reduction in congestion could also improve air quality, as congested streets are a major driver in air pollution [15].
Current Challenges
Autonomous technology still has numerous challenges to overcome until it’s ready for mass market adoption, however. This has become especially apparent in the last few years. Five years ago, truly autonomous driving technology was widely predicted to become market ready by 2020 [16]. This has not come true, however. Although the technology has certainly become more sophisticated in the last five years, we’re still some time away from a fully automated automotive future. As of 2021, most of the autonomous car industry has pushed back their timeline of when autonomous driving technology will be commercially available [17]. In recent studies, at least partly automated vehicles are now predicted to make up between 30% and 45% of all cars in Germany by 2050 [18]. Level five autonomy is predicted to be market ready by 2030 at the earliest [19]. All in all, self-driving technology is expected to take longer than expected until it’s market ready.
There are several reasons as to why this is the case. Firstly, and most importantly, the industry had misjudged the complexity of self-driving technology and its edge cases. As the technology has been developed further and more test drives were performed, the true complexity of it has become more apparent [17]. This is most apparent in edge cases, in which two very similar situations might call for widely different reactions. For example, an autonomous car will have to differentiate between harmless and dangerous obstacles (e.g. a flock of birds compared to leaves in the wind. While modern autonomous vehicles are able to drive safely in most conditions, interpreting these edge cases is more difficult than expected [21]. The complexity of autonomous driving technology has also been made apparent by fatal accidents involving autonomous cars. Some experts believe that specific incidents, like the fatal crash in Phoenix Arizona in 2018 reset expectations within the industry [23].
Another limiting factor for driverless technology is consumer perception. As recent consumer studies have shown, the majority of drivers in Germany as well as the United States do not trust autonomous cars yet. In a 2021 study in Germany, 30% of respondents claimed they would not want to use a fully autonomous vehicle. Among the respondents, there were significant differences between gender and age groups. For instance, the willingness of using autonomous vehicles was almost twice as high among men compared to women [24]. This could have multiple reasons. For once, trust might be low as a result of the recent fatal accidents involving autonomous cars. Also, knowledge about the technology might not be very widespread, which could be indicated by the high proportion of respondents who claimed they don’t have an opinion on autonomous technology. According to a study conducted by the American Automobile Association, only 22% of American car owners felt like manufacturers should focus on autonomous cars while 80% would like to see improvements to current vehicle safety systems [25]. Overall, trust in the technology is still relatively low, meaning it has yet to prove itself to the general public.
Lastly, driverless technology has significant legal and ethical implications. For instance, as cars solely rely on sensors and artificial intelligence to navigate, the question of who is responsible for accidents or other incidents arises [26]. This question is especially relevant in vehicles, where passengers cannot intervene in its behavior anymore.
Furthermore, the decision framework needed for a car to make consistent and reliable decisions is complex and ambiguous. As a result, manufacturers will need to create ethical frameworks for how the car behaves in critical situations (e.g. when a crash is unavoidable). For instance, vehicles will need to judge and decide whose safety and well-being is prioritized. In extreme cases, the vehicle would essentially have to decide over whose survival is more important [22].
One aspect of making autonomous cars safe and reliable is so-called vehicle-to-everything (V2x) communication. This technology would allow cars to communicate with other traffic participants. By 2020 the implementation of V2x was still pending [20].
Outlook
Autonomous cars are one of the biggest shifts in the automotive industry in the last century. The technology promises to alleviate stress and improve quality of life by giving commuters and other traffic participants more time to work or relax. At the same time, smart networking between cars and other traffic participants can make roads safer, less congested and as a result help in reducing pollution. Electric mobility combined with autonomous vehicles can further amplify this effect. If autonomous technology becomes market ready and available, new business models like robo-taxis could also decrease car ownership.
We see autonomous vehicles in passenger transport as an important innovation. In conjunction with new business models and electric mobility, they have the potential to improve car traffic in numerous ways. The true potential of the technology depends on multiple factors, however. For once, it remains to be seen how quickly the technology is ready for the market. Secondly, it depends on how quickly adoption of autonomous vehicles will progress. Customer acceptance plays a role in this context, as a large proportion of people are currently still skeptical about autonomous driving. If these requirements are met, autonomous technology can improve our car traffic significantly.
Sources
[1] Dungs, J., Duwe, D., Schmidt, A., Herrmann, F., & Stegmüller, S. (2016). The Value of Time: Nutzerbezogene Service-Potenziale durch autonomes Fahren. Fraunhofer. https://blog.iao.fraunhofer.de/images/blog/studie-value_of_time.pdf
[2] SAE International. (2018, December 11). SAE J3016 Levels of driving automation. https://www.sae.org/news/press-room/2018/12/sae-international-releases-updated-visual-chart-for-its-“levels-of-driving-automation”-standard-for-self-driving-vehicles
[3] Gessner, D. (2020, July 22). Experts say we're decades from fully autonomous cars. Here's why. Business Insider. https://www.businessinsider.com/self-driving-cars-fully-autonomous-vehicles-future-prediction-timeline-2019-8
[4] Krafcik, J. (2020, October 8). Waymo is opening its fully driverless service to the general public in Phoenix. Waymo. https://blog.waymo.com/2020/10/waymo-is-opening-its-fully-driverless.html
[5] California Department of Motor Vehicles. (2021). California DMV Tesla Robo-Taxi / FSD Notes. https://www.plainsite.org/documents/28jcs0/california-dmv-tesla-robotaxi--fsd-notes/
[6] Statista. (2021, February 15). Infographic: The Self-Driving Car Companies Going The Distance. https://www.statista.com/chart/17144/test-miles-and-reportable-miles-per-disengagement/
[7] Etherington, D. (2020, June 26). Amazon to acquire autonomous driving startup Zoox. TechCrunch. https://techcrunch.com/2020/06/26/amazon-to-acquire-autonomous-driving-startup-zoox/
[8] Nellis, S., Shirouzu, N., & Lienert, P. (2020, December 21). Exclusive: Apple targets car production by 2024 and eyes 'next level' battery technology. Reuters. https://www.reuters.com/article/us-apple-autos-exclusive-idUSKBN28V2PY
[9] Manyika, J., Chui, M., Bughin, J., Dobbs, R., Bisson, P., & Marrs, A. (2013). 7. Disruptive technologies: Advances that will transform life, business, and the global economy. McKinsey. https://www.mckinsey.com/~/media/mckinsey/business%20functions/mckinsey%20digital/our%20insights/disruptive%20technologies/mgi_disruptive_technologies_full_report_may2013.pdf
[10] Oldenburg, R. (1999). 15. The great good place: Cafes, coffee shops, bookstores, bars, hair salons, and other hangouts at the heart of a community. Da Capo Press.
[11] Statistisches Bundesamt. (2021). Fehlverhalten der Fahrer bei Unfällen mit Personenschaden. https://www.destatis.de/DE/Themen/Gesellschaft-Umwelt/Verkehrsunfaelle/Tabellen/fehlverhalten-fahrzeugfuehrer.html
[12] Eno Center for Transportation. (2013). 4-8. Preparing a Nation for Autonomous Vehicles: Opportunities, Barriers and Policy Recommendations. https://www.enotrans.org/wp-content/uploads/AV-paper.pdf
[13] Ambadipudi, A., Heineke, K., Kampshoff, P., & Shao, E. (2017, September). 9-10. Gauging the disruptive power of robots-taxis in autonomous driving. McKinsey & Company. https://www.mckinsey.com/~/media/McKinsey/Industries/Automotive%20and%20Assembly/Our%20Insights/Gauging%20the%20disruptive%20power%20of%20robo%20taxis%20in%20autonomous%20driving/Gauging-the-disruptive-power-of-robo-taxis-in-autonomous-driving.pdf
[14] Geistefeldt, J., & Lohoff, J. (2011). 14. Stausituation auf den Autobahnen in Nordrhein-Westfalen. Ruhr Universität Bochum. https://www.vm.nrw.de/presse/_container_presse/StausituationNRW_Schlussbericht_Endfassung-Presse.pdf
[15] Zhang, K., & Batterman, S. (2013). Air pollution and health risks due to vehicle traffic. Science of The Total Environment, 450-451, 307-316. https://doi.org/10.1016/j.scitotenv.2013.01.074
[16] Warren, T. (2016, November 1). The US Secretary of transportation on a driverless America. The Verge. https://www.theverge.com/a/verge-2021/secretary-anthony-foxx
[17] Downs, G., & Yi, H. (2021, May 20). Beyond Tesla: Driverless Startups Promise Next-Level Autonomous Vehicles [Video]. YouTube. https://www.youtube.com/watch?v=UdOxt11ofjQ
[18] Altenburg, S., Kienzler, H., & Auf der Maur, A. (2018). 22. Einführung von Automatisierungsfunktionen in der Pkw-Flotte. Prognos. https://www.prognos.com/sites/default/files/2021-01/adac_automatisiertes_fahren_endbericht_final.pdf
[19] McKinsey & Company. (2021). Autonomous Driving - Overview. https://www.mckinsey.com/features/mckinsey-center-for-future-mobility/overview/autonomous-driving
[20] Hawkins, A. J. (2020, April 23). The auto industry’s fight with the FCC over ‘vehicle-to-everything’ communication is heating up. The Verge. https://www.theverge.com/2020/4/23/21233085/v2x-vehicle-to-everything-fcc-safety-spectrum-airwaves-wifi
[21] Adams, E. (2020, September 25). Why we’re still years away from having self-driving cars. Vox. https://www.vox.com/recode/2020/9/25/21456421/why-self-driving-cars-autonomous-still-years-away
[22] Maurer, M., Gerdes, J. C., Lenz, B., & Winner, H. (2015). 69-82. Autonomes Fahren: Technische, rechtliche und gesellschaftliche Aspekte. Springer Vieweg.
[23] Bates, I. C. (2020, May 12). This was supposed to be the year driverless cars went mainstream. The New York Times. https://www.nytimes.com/2020/05/12/technology/self-driving-cars-coronavirus.html
[24] Statista. (2021, August). Interesse deutscher Autofahrer an autonomen Fahrzeugen 2021. https://de.statista.com/statistik/daten/studie/270596/umfrage/interesse-an-autonomen-fahrzeugen/
[25] Edmonds, E. (2021, February 25). Today’s Vehicle Technology Must Walk So Self-Driving Cars Can Run. American Automotive Association. https://newsroom.aaa.com/2021/02/aaa-todays-vehicle-technology-must-walk-so-self-driving-cars-can-run/
[26] Laing, K. (2021, May 20). Self-Driving Cars Pose Crucial Question: Who to Blame in a Crash. Bloomberg. https://www.bloomberg.com/news/articles/2021-05-20/self-driving-cars-pose-crucial-question-who-to-blame-in-a-crash
[27] Gurman, M. (2021, November 18). Apple Accelerates Work on Car Project, Aiming for Fully Autonomous Vehicle. Bloomberg. https://www.bloomberg.com/news/articles/2021-11-18/apple-accelerates-work-on-car-aims-for-fully-autonomous-vehicle