FORECASTING AMERICANS’ LONG-TERM ADOPTION OF CONNECTED AND AUTONOMOUS VEHICLE TECHNOLOGIES

The introduction highlights the significant advancements in automotive technology, particularly the emergence of connected and autonomous vehicles (CAVs). While CAVs offer potential benefits like reduced crashes and congestion, they also raise concerns about safety, security, privacy, and economic transition impacts. The study addresses the uncertainty surrounding CAV adoption by developing a simulation-based framework that considers both demand-side (WTP) and supply-side (technology prices) factors, as well as government regulations. Existing studies are criticized for their limited consideration of demand-side factors, government regulations, and the adoption of specific CAV technologies (Level 1 and 2). This research aims to fill these gaps and provide more robust adoption rate forecasts.

The literature review examines existing studies on CAV adoption rate forecasting. It highlights the varied predictions from researchers, private enterprises, and industry enthusiasts, criticizing the reliance on extrapolating trends from past technologies, expert opinions, or supply-side variables, without adequate consideration of underlying assumptions and demand-side factors. The review points out the lack of attention to consumers' WTP, vehicle transaction decisions, and government regulations in previous works, along with the absence of mechanisms for anticipating the adoption of specific Level 1 and 2 automation technologies and vehicle connectivity. This study is positioned as the first to forecast long-term CAV fleet evolution, considering consumer WTP, technology prices, and NHTSA regulations.

The methodology involved a multi-stage process. First, a survey was conducted using Qualtrics, collecting data from 2,167 Americans on their vehicle ownership, transaction decisions, WTP for CAV technologies, and opinions on CAVs. The survey data was then weighted to account for sample biases using the 2013 American Community Survey data. Respondents' home addresses were geocoded to incorporate built-environment factors into the analysis. Logit models were developed to understand the impact of demographics and built-environment variables on vehicle transaction decisions. A simulation-based framework, incorporating the logit models and Monte Carlo simulations, was used to forecast annual technology adoption rates from 2015 to 2045. The framework simulates households’ annual decisions: selling, buying, replacing, adding technology, or doing nothing. The probabilities of these decisions were determined using a multinomial logit (MNL) model estimated in BIOGEME. The framework considers the acquisition of new versus used vehicles and the associated costs of adding technologies. The impact of NHTSA regulations mandating ESC and connectivity was also integrated into the simulation.

The key findings are presented across several aspects. Regarding WTP, the average WTP (among respondents with non-zero WTP) for connectivity, Level 3, and Level 4 automation was $110, $5,551, and $14,589, respectively. For Level 1 and 2 technologies, interest and WTP varied significantly, with blind-spot monitoring and emergency automatic braking being the most appealing. The average WTP for most Level 1 and 2 technologies was lower than the projected future price, suggesting potential affordability challenges. Concerning opinions, most respondents perceived themselves as good drivers, enjoyed driving, and preferred to wait before adopting new technologies. While many saw AVs as useful, a significant portion expressed fear. Trust in technology companies and manufacturers influenced adoption opinions. The simulation results demonstrated significant differences in long-term adoption rates across eight scenarios, highlighting the impact of technology pricing, WTP, and regulations. Scenarios with 10% annual price drops or WTP increases showed substantially higher adoption rates for all technologies by 2045. Government regulations on ESC and connectivity significantly increased their adoption rates. The interplay between WTP, technology pricing, and regulations shaped the adoption of Level 3 and 4 automation; increasing WTP and price drops increased Level 4 adoption while decreasing Level 3 adoption due to the simulation's hierarchical framework, which prioritized Level 4. Adoption rates of Level 1 technologies varied depending on the scenario but generally increased with higher WTP and lower prices. The study concludes that significant changes in WTP or rapid price reductions are needed for widespread CAV adoption.

The findings highlight the critical role of consumer behavior (WTP) and technological advancements (price) in influencing CAV adoption. The study successfully demonstrates how a simulation framework incorporating these factors, alongside governmental regulations, produces more nuanced and realistic adoption rate forecasts than previous studies. The strong impact of regulations on the adoption of ESC and connectivity underscores the potential of policy interventions in accelerating technological uptake. The results suggest that technological cost reduction and increasing consumer confidence are essential for wider adoption of advanced automation technologies. The study provides valuable insights for policymakers, industry stakeholders, and researchers involved in shaping the future of the automotive industry.

This study offers a robust simulation-based framework for forecasting long-term CAV adoption rates. The results demonstrate the importance of considering consumer WTP, technology prices, and government regulations when predicting technological diffusion. The study's findings offer critical insights for policymakers and industry stakeholders. Future research could incorporate household evolution models to improve WTP estimations and include shared autonomous vehicles (SAVs) to better capture their potential impact on vehicle ownership and adoption patterns.

The study's limitations include the reliance on a 2015 survey, which may not fully reflect evolving consumer preferences. The model's assumptions, such as constant annual WTP increases and price reductions, could be refined with more dynamic data and incorporating behavioral economic models. Furthermore, the study does not fully account for potential unforeseen events (like a major safety incident) which may impact adoption rates. The hierarchical approach to assigning Level 3 and Level 4 automation may influence the results, while the exclusion of factors such as infrastructure limitations and charging infrastructure for electric vehicles also represents limitations.