Automaker strategies and AV diffusion scenarios

In this segment the potential AV diffusion scenarios and automaker strategies are discussed to evaluate when fully autonomous vehicles could become available to the public and at what rate they could gain popularity.

Automakers have adopted different strategies in their development efforts of AVs.

Most modern vehicles have driver assistance systems in them as standard features, and SAE level 2 systems are available even in affordable car segments (Horaczek 2018). Audi’s flagship model A8 became the world’s first production car to achieve true SAE level 3 autonomy in 2017, while the rest of the manufacturers are either still developing systems capable of conditional automation or skipping it altogether

(Taylor 2017). Automakers such as Toyota, Ford and Volvo have deemed SAE level 3 automation unsafe and unfeasible, and have set their aim straight to high automation (Robotics Law Journal 2017). Nissan, Daimler and General Motors have set a goal to rollout their versions of AVs in the year 2020, although little information has been released stating the degree of autonomy of these vehicles (Beiker 2016).

The wide range of automaker strategies can be credited to the fact that the industry is unsettled on what will be the dominant design of driving automation. Referring back to Abernathy and Utterback’s phases of industrial innovation, the development of AVs could be described as still being in the fluid phase. Whether conditional and high automation are competing designs is not however a straightforward question.

Even though full automation may in the end phase out conditional automation the same way AVs could replace the human driver, SAE level 3 can potentially have several decades of vitality in it before SAE levels 4 and 5 take over (Bierstedt et al 2014; Milakis et al 2017a). Conditional automation can also prove important in molding legislation and public perception of AVs in preparation for further automation levels (Gasser 2016; Winkle 2016).

Table 4. Diffusion estimates of AV levels (Nieuwenhuijsen et al 2018) Introduction Market

2017–2020 70% in 2020 Underwood (2014), Rangarajan and Dunoyer (2014), Juliussen and Carlson (2014)

Level 4, high 2018–2024 Highway, some urban streets before

Bierstedt et al (2014), Juliussen &

Carlson (2014), Rangarajan &

Dunoyer (2014), Underwood (2014), Litman (2015), Milakis et al (2017a)

The synthesis shown in Table 4 gives an early understanding of the possible diffusion of AVs, and its scale of time. Currently there is no consensus in academia over the prospective innovation diffusion of the various levels of AVs, and even the terminology used across the studies is different (Nieuwenhuijsen et al 2018). Due to the long range of the forecasts, any growth scenarios for AVs at this point are highly speculative.

AV academia estimated the introduction of SAE level 3 automation correctly, but wildly missed the estimations for its market penetration. In the case of full automation, scholars and industry experts estimate that the market introduction could take place any time between 2025 and 2045, while the public polls place introduction around the year 2030 (De Winter et al 2014; Kyriakidis et al 2015;

Milakis et al 2017a). In the most favorable scenarios fully autonomous vehicles are introduced in the second half of 2020 and could make up for more than half of the entire vehicle fleet by the 2050s (Bierstedt et al 2014; Litman 2015; Milakis et al 2017a). Talebian and Mishra (2018) estimate using IDT and agent-based modelling that AVs could reach market saturation by 2050, but this would require annual price reductions for AVs to average 15 to 20 percent.

Figure 13. Estimated deployment scenario for SAE level 5 AVs (Litman 2018;

Nieuwenhuijsen et al 2018)

Figure 13 shows a hypothetical deployment pattern for fully autonomous vehicles based on innovation diffusion literature synthesis by Nieuwenhuijsen et al (2018), which is drawn upon the innovation S-curve illustration by Litman (2018). In this scenario the most rapid phase of diffusion of SAE level 5 AVs could take place between the early 2040s and late 2050s given that the technology becomes commercially available in the first half of the 2030s. While this could be considered

“a middle of the line” estimation when looking at the spectrum of potential outcomes suggested by various AV diffusion studies, there are too many uncertainties at this point to accurately evaluate what is most likely going to happen.

What the research community is more congruent about are the factors which support technological development and market proliferation of AVs. This has become to be known as the “AV in bloom” -scenario, in which the technological development of AVs is strong, AVs are supported by legislators and public perception of consumers is good (Fagnant & Kockelman 2015; Milakis et al 2017a;

Nieuwenhuijsen et al 2018). Nieuwenhuijsen et al (2018) found that policy instruments which support knowledge transfer and creation of external research funds are the most effective measures at accelerating market take-up of high levels of automation.

Even in the most optimistic scenarios there are tangible factors which greatly slow down the AV diffusion process. High purchase prices and rapid value depreciation of new cars, long lifespans, and a vibrant secondhand market decelerate the diffusion of new vehicle technologies as they reduce the demand for new vehicles overall (Berkovec 1985; Garsten 2018; Litman 2018). It is therefore unlikely that the most enthusiastic AV diffusion scenarios materialize (Litman 2018) All in all, the accuracy of the diffusion forecasts will greatly improve once high and full automation levels become commercially available and actual early sales data can be utilized (Bass 2004; Massiani & Hogs 2015; Cooper & Gutowski 2018).