New technologies are emerging in the private vehicle market. Conventional propulsion systems are set to be replaced by alternative, more environment-friendly ones (e.g., electric vehicles), and certain new features, like autonomous driving, will possibly change the way private cars are employed. To assess the impact of such technologies, one must estimate how often and for which trips these vehicle types will be used. Another issue is the exact localization of certain vehicle types on the network, to assess environmental effects and identify where specific roadside infrastructure (e.g., charging stations) will be required. This paper presents four approaches to forecasting car usage by vehicle type using a macroscopic travel demand model in combination with a vehicle fleet or technology diffusion model. Integrating the two types of models requires tools ranging from assumptions and extrapolation of empirical data to synthetic or incremental discrete choice models. The approaches are employed in a case study forecasting travel demand using privately owned autonomous vehicles (AVs) in Germany in 2030. Despite identical input data, the estimated proportion of vehicle miles traveled (VMT) using AVs varies between 11% and 23% of overall car VMT, depending on the approach chosen. The reasons for this variation in results are investigated and some recommendations are given. To avoid the difficulties of fitting a synthetic model to observed data and to increase the accuracy of the results, the recommendation is to formulate the vehicle type choice as an incremental model added to the travel demand model.