Simulating long-term emissions from private automated vehicles under climate policies
Introduction
Widespread uptake of fully automated vehicles (AVs) may significantly transform the transport system in the coming years (Bansal and Kockelman, 2017). In this study we simulate the potential impacts of AV availability (starting in 2025) on light-duty vehicle greenhouse gas (GHG) emissions out to 2035. We use the term AVs (short for fully automated vehicles) to refer to vehicles with an automation level of 4 (or higher) as per the definition used by the US National Highway and Traffic Safety Administration (2022). Others might call these autonomous or self-driving vehicles—but we use the term “fully automated”. We assume that any drivetrain (powered by gasoline or electricity) can be fully automated with the use of LIDARs and cameras and other required software.
Apart from the wide-ranging impacts on road safety, automobility, and travel behaviour, large scale AV usage could also have significant energy and GHG impacts (Wadud et al, 2016). Of course, there is enormous uncertainty regarding AV deployment, including availability and consumer adoption. Of particular interest in this paper, there are a variety of ways that AVs can impact GHG emissions, including changes in driving efficiency (or inversely, energy intensity), vehicle sizing, traffic congestion, and vehicle travel demand for existing and new user groups (vehicle km travelled or VKT). The collective understanding of these future trends and impacts is still in its infancy (Harb et al, 2021).
Simultaneously, governments across the world are implementing climate policies to reduce GHG emissions from the road transport sector. For example, our case study of Canada has implemented a mix of climate policies to achieve net zero emissions by 2050, including carbon pricing and regulations such as a low-carbon fuel standard and vehicle emissions standard (Government of Canada, 2021a). In line with this longer-term objective, many jurisdictions have announced goals for light-duty vehicle sales to be 100% zero-emissions vehicles (ZEVs) by 2035, such as in Canada, the UK, the Netherlands, Norway, and the US state of California. A subset of these regions (in the US and Canada) have translated these goals into a binding ZEV sales mandate (Axsen et al., 2022). This push towards deep GHG reduction goals and climate policies leaves many unanswered questions about how AVs might interact with climate policies and ZEVs, and about their resulting GHG impacts.
Within this context, we presently explore the following research questions, focusing on the case of Canada’s light-duty vehicle sector:
- 1.
If AVs become available for sale, what sales rates are expected with and without climate policy (carbon pricing and regulation)?
- 2.
How does AV availability impact ZEV sales (starting in 2025)?
- 3.
What are the GHG emissions impacts of AVs with and without climate policy, when considering potential changes to efficiency and vehicle travel (VKT)?
In this study, we use AUM, a technology adoption model that simulates the uptake of light duty vehicles in Canada—which we adapt to represent AV technology. By ‘simulating’, we mean projecting future consumer adoption of new and emerging technologies under a certain set of behaviour, technology, and policy assumptions. In this case we study the uptake of AVs, ZEVs, and vehicles that combine both features. Technology adoption models are often associated with the larger category of energy-economy models. Within the literature on technology adoption and energy-economy models, some focus only on conventional internal combustion engine vehicles (e.g. Small, 2012), and some others focus on both internal combustion engine vehicles and ZEVs (e.g. Fox et al., 2017, Wolinetz and Axsen, 2017). ZEVs are vehicles that have the potential for zero tailpipe emissions, namely plug-in hybrid (PHEVs), battery-electric (BEVs), along with fuel-cell electric vehicles (FCEVs). Our study extends this literature space by including AVs, in addition to including internal combustion engine vehicles and ZEVs. Though, since the uptake of light-duty fuel cell vehicles globally remains very low, we exclude this drivetrain technology in our current study.
As discussed in greater detail in the literature review section, this study is novel on several accounts. First, in contrast to the few previous attempts of estimating long-term adoption of AVs (e.g., Bansal and Kockelman, 2017, Talebian and Mishra, 2018), we use a technology adoption model which includes behaviourally realistic representations of consumers, including dynamic consumer preferences, and endogenous representation of automakers including technology learning. Second, this model simulates demand- and supply-side dynamics to allow us to endogenously break down the GHG impacts of AVs into several constituent categories, such as efficiency (or inversely energy intensity), VKT response, and interactions with ZEV adoption. In contrast, previous literature estimating the GHG impacts of AVs tends to rely on exogenous assumptions and calculations (e.g., Wadud et al., 2016, Milakis et al., 2017), or else uses optimization models that do not represent realistic consumer preferences (e.g., Jones and Leibowicz, 2019, Brown and Dodder, 2019, Sheppard et al., 2021). Third, our application includes a novel analysis of the potential interactions between ZEV supportive policies and AV adoption, which is missing from existing literature.
This current study focuses on private AVs. Some researchers argue that future scenarios with widespread usage of shared AVs (SAVs or “robo-taxis” in some form) are feasible and could have positive impacts on society, including dramatic reductions in GHG emissions (Sperling et al., 2018, Vilaça et al., 2022). However, emerging empirical evidence suggests that the majority of consumer segments prefer private AVs over SAVs (Haboucha et al., 2017, Nazari et al., 2018, Saleh and Hatzopoulou, 2020, Saeed et al., 2020), and researchers agree that there is great uncertainty over whether (and under what conditions) consumers will choose private AVs or SAVs (Wadud and Chintakayala, 2021). While a future of SAVs that displace private vehicle ownership is technically possible, we presently focus on the study of AV uptake in the private light-duty vehicle sector.
In short, our present research objectives are to simulate the adoption of privately owned AVs in the Canadian light duty vehicle sector, using a model that represents consumer behaviour, supplier decision-making, and climate policy. Using the time horizon to 2035, we simulate AV new sales each year and changes in light-duty vehicle stock, and estimate the GHG impacts of AV adoption. To simulate the effects of policy on GHG emissions from AV deployment, we consider scenarios with baseline policies, and with added “strong” carbon pricing or ZEV regulation. We also represent uncertainty in numerous parameters, and report the sensitivity of results to these parameters.
The paper is structured as follows. Section 2 reviews the literature on AVs, Section 3 describes the model used for the study, Section 4 presents the results, and Section 5 concludes with a discussion on the results.
Section snippets
Literature review
Adapting the classification from Zhang and Zhang (2022), AV-related literature can be categorized according to at least four sub-themes (though with significant overlap), namely:
- i.
survey studies assessing the consumers’ willingness to pay for AVs, or factors affecting consumers’ acceptance of AVs;
- ii.
studies on the impact of shared AVs on urban road traffic, trip generation, mode and route choice and road networks;
- iii.
studies on (long-term) adoption and diffusion of AV technology; and
- iv.
studies on energy
The AUtomaker-consumer model (AUM)
We use the AUtomaker-consumer Model (AUM) to simulate the impacts of AVs on Canada’s light-duty vehicle sector. AUM is a technology adoption model which simulates long-term (from 2020 to 2035) projections of new light-duty vehicle sales (including ZEVs), total light-duty vehicle stock, and the corresponding GHG emissions. AUM is unique in that it includes behaviorally-realistic consumers (also accounting for consumer preferences, rather than just financial cost considerations in technology
AV market share and its interaction with ZEV uptake
Fig. 3 depicts median new AV market share under the three policy scenarios (Baseline, Baseline + Tax, Baseline + ZEV), where the uncertainty in market share estimates is depicted by the optimistic and pessimistic values in Table 9. Under the Baseline scenario with median values, new AV market share (starting in 2025) is expected to be 12.7% in 2025 and rising to 24.1% by 2035.
Relative to the Baseline, the market share of AVs is marginally lower under the climate policy scenarios (Baseline + Tax
Discussion and conclusions
In this study, we attempt to answer the following research questions: i) If AVs become available for sale, what sales rates are expected with and without climate policy (carbon pricing and regulation)? ii) How does AV availability impact ZEV sales? iii) What are the GHG emissions impacts of AVs with and without climate policy?
For our research we use AUM, a light-duty vehicle technology adoption model which includes dynamic consumer preferences, endogenous technology learning, consumer
Author contribution statement
The authors confirm contribution to the paper as follows: study conception and design: Chandan Bhardwaj; Jonn Axsen; Curran Crawford; data collection (literature review), model development and analysis: Chandan Bhardwaj; draft manuscript preparation: Chandan Bhardwaj, Jonn Axsen, Curran Crawford; All authors reviewed the results and approved the final version of the manuscript.
Declaration of Competing Interest
Funding was provided by the Community Trust Endowment Fund (CTEF) of Simon Fraser University, and the Social Sciences and Humanities Research Council (SSHRC) of Canada. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References (96)
- et al.
Capturing the behavioural determinants behind the adoption of autonomous vehicles: Conceptual frameworks and measurement models to predict public transport, sharing and ownership trends of self-driving cars
Transport. Res. F: Traffic Psychol. Behav.
(2019) - et al.
A critical review and assessment of Eco-Driving policy & technology: Benefits & limitations
Transp. Policy
(2014) - et al.
Forecasting Americans’ long-term adoption of connected and autonomous vehicle technologies
Transp. Res. A Policy Pract.
(2017) - et al.
Endogenizing R&D and market experience in the “bottom-up” energy-systems ERIS model
Technovation
(2004) - et al.
Uncovering early adopter’s perceptions and purchase intentions of automated vehicles: Insights from early adopters of electric vehicles in California
Transport. Res. F: Traffic Psychol. Behav.
(2019) - et al.
Simulating automakers’ response to zero emissions vehicle regulation
Transp. Res. Part D: Transp. Environ.
(2021) - et al.
Which “second-best” climate policies are best? Simulating cost-effective policy mixes for passenger vehicles
Resour. Energy Econ.
(2022) - et al.
Modeling the uptake of plug-in vehicles in a heterogeneous car market using a consumer segmentation approach
Transp. Res. A Policy Pract.
(2017) - et al.
Energy and emissions implications of automated vehicles in the US energy system
Transp. Res. Part D: Transp. Environ.
(2019) - et al.
Are consumers willing to pay to let cars drive for them? Analyzing response to autonomous vehicles
Transport. Res. Part C: Emerg. Technol.
(2017)