Modeling the Usage of Future Vehicle Types

Abstract

New technologies are emerging in the private vehicle market. On the one hand, conventional propulsion systems are set to be replaced by alternative, more environment-friendly ones (e.g. electric vehicles). One the other hand, certain new features, like vehicle automation, will possibly change the way private cars are used. In order to assess the impact of such technologies, it is necessary to estimate how often and for which trips these vehicles will be used. Another important issue is the exact localization of certain types of vehicles on the street network, in order to determine where specific roadside infrastructure will be required. This paper presents several approaches to estimate the usage of specific types of private vehicles. One approach is to use empirical trip data to create homogenous categories of trips according to trip length, purpose etc. For these trip categories, the usage of vehicle types is assumed to remain unchanged and to be influenced only by certain vehicle attributes like size, age etc., meaning that new vehicle types will be used just like current ones if they are grouped in the same category. A different approach is to use discrete choice modeling to calculate the vehicle shares. The challenge using this approach is to identify the relevant variables and estimate a model consistent with input data.