Modeling the market penetration of automated vehicles with an integrated choice-based diffusion approach

In the past decade, transportation researchers have predicted Automated Vehicle (AV) preferences using discrete choice models, including logit and probit models. These studies rely on stated preferences on AV adoption which only provides an adoption scenario at a single point in time. This limitation can be overcome by following a diffusion-based approach known as the Bass Model using sales data on the product. In this study, we use existing sales data on Level 2 AVs also known as advanced driver assistance systems (ADAS) for prediction of mass adoption of Level 3 AVs. This study is the first to model AV mode choice combining discrete choice theory with a diffusion model. Stated and revealed preference data from Puget Sound Travel Survey, and a Nested Logit Model are used to obtain the market penetration of AV travel modes. The market shares are then fed into the diffusion model for the prediction of the number of adopters for automated transit, automated taxi and privately owned autonomous cars. Our findings reveal that it will take 20 years to reach full market potential for Level 3 automated cars and buses, when there is a 0.53 percent annual decrease in the price of the ADAS technology. By quantifying the timing and mode-specific dynamics of AV adoption under different cost scenarios, this study may help shape policies in the future related to the penetration of AVs, specifically pricing policies, and regulatory strategies for varying AV modes.