查看文章

In this study we investigate how partially automated vehicles (Tesla electric vehicles with “Autopilot”) are used, including how often automation is used, on what roads, in what weather, and in what traffic conditions. We use a latent class model to identify heterogenous classes of autopilot users, then we use a multinomial logistic regression model to understand the relationship between each latent class and several independent variables, including socio-demographics and vehicle miles travelled (VMT). The latent class model revealed four latent classes: very frequent users, who use it most frequently; frequent users, who use automation frequently in clear weather and on freeways; semi-frequent users who use it for less than half their trips and only on freeways, in clear weather, when there is no traffic; and infrequent users, who use it the least often and only in clear weather, on freeways, when there is no traffic. The …