Michael Heilig

Hybrid Modeling Approach of Car Uses in Germany on Basis of Empirical Data with Different Granularities

The use of private cars in Germany has not yet been analyzed from a longitudinal perspective: most travel surveys consider only a single day. Daily car usage is not identical over a given period because car owners use their vehicles for daily routines (e.g., commuting) as well as for infrequent events, such as holiday trips. Another problem of short-period surveys is that they underestimate the share of cars used for long-distance travel. The current work may help to improve the reliability and realism of statements about the extent to which German cars could be replaced by electric vehicles. The authors developed a hybrid modeling approach that aims to obtain car mileage per day for a full year. This approach is based on empirical data with different granularities. Input data are derived from the annually conducted German Mobility Panel, including a survey of fuel consumption and odometer readings, and the long-distance travel survey INVERMO. The study showed that 13.1% of the modeled German private car fleet never exceeded 100 km/day during a full year. Furthermore, cars were driven more than 100 km on 13.3 days/year on average. Mainly used cars (first cars) of a household were used for longer distances rather than second cars. A comparison of average mobility figures from the model approach with the Mobility in Germany national travel survey showed the model results as reliable and realistic.

Modeling Week Activity Schedules for Travel Demand Models

Activity schedules are an important input for travel demand models. This paper presents a model to generate activity schedules for one week. The approach, called actiTopp, is based on the concept of utility-based regression models and stepwise modeling. In contrast to most of the existing models, actiTopp covers the time period of one week. Few models have covered one week; thus, the activity generation approach of this simulation period is rare. Analysis of weekly activity behavior shows stability between different days (e.g., working durations). Hence, the model explicitly takes these aspects into account, for example, by defining time budgets to spread durations within the week. For model estimation, the study used data from the German Mobility Panel (MOP). This annual survey collects representative data on the travel behavior of the German population. The data from 2004–2013 provide more than 17,500 activity schedules for one week, with more than 450,000 activities. Selected results are shown for the model application to 2014 MOP data, which the study used for validation purposes. The mean value of activities per person and week show a difference of 0.3 activity. To evaluate the model, the study used Kolmogorov-Smirnov tests with a significance level of α = 0.001. For the activity type distribution of the 2014 sample, the analysis could not reject the null hypothesis of equality of the distribution of the model and the survey data at this significance level.

Large-Scale Application of a Combined Destination and Mode Choice Model Estimated with Mixed Stated and Revealed Preference Data

The diffusion of new modes of transportation, such as carsharing and electric vehicles, makes it necessary to consider them along with traditional modes in travel demand modeling. However, there are two main challenges for transportation modelers. First, the new modes’ low share of usage leads to a lack of reliable revealed preference data for model estimation. Stated preference survey data are a promising and well-established approach to close this gap. Second, the state-of-the-art model approaches are sometimes stretched to their limits in large-scale applications. This research developed a combined destination and mode choice model to consider these new modes in the agent-based travel demand model mobiTopp. Mixed revealed and stated preference data were used, and new modes (carsharing, bikesharing, and electric bicycles) were added to the mode choice set. This paper presents both challenges of the modeling process, mainly caused by large-scale application, and the results of the new combined model, which are as good as those of the former sequential model although it also takes the new modes into consideration.

Modelling the Weekly Electricity Demand Caused by Electric Cars

Abstract As transport is one of the big sources of carbon dioxide emissions, it is natural to seek for solutions reducing the carbon dioxide emissions in transport as well. Replacing cars powered by a combustion engine by battery electric vehicles may be one measure to achieve this goal, at least as long as the electricity consumed by these cars is produced carbon neutral or in a low-carbon manner. In Germany, the Federal Government aims at a stock of one million electric vehicles in the year 2020. This goal is very ambitious, since customers are reluctant to buy battery electric cars, probably most of all due to their limited range. A possible solution to the limited range problem is the use of Plug-in Hybrid Electric Vehicles or Extended Range Electric Vehicles (EREV), combining an electric battery with a combustion engine or a generator. These solutions overcome the range limitations while at the same time allowing driving on electric power for the majority of the total mileage. In this paper, we analyse the effects of an increased use of EREVs and battery electric vehicles using the travel demand model mobiTopp. For three scenarios with different rates of market penetration of electric vehicles, the travel demand and car usage is simulated over a simulation period of one week. The results show, that for 65 up to 70 percent of the mileage, EREVs can be driven in battery-only mode, demonstrating the usefulness of the EREV concept and indicating a substantial potential for the reduction of carbon dioxide emissions. The results, however, also show that with an uncontrolled charging strategy, i. e. every car recharges immediately after accessing a charging location, the peaks of electricity demand for charging the electric cars occurs when the general electricity demand is already high. During these periods, additional electricity demand is typically covered by gas-fuelled power plants, thus using fossil fuels. Therefore, the concept of introducing electric vehicles in order to reduce total carbon dioxide emissions can only succeed if combined with intelligent charging strategies.

Microscopic Demand Modeling of Urban and Regional Commercial Transport

Abstract Commercial transport is an intrinsic part of the evaluation of traffic volumes. However, it is often limited to freight transport, and while this is a significant element, it disregards the share of trips contributed by plumbers, electricians, care services, and the like. These businesses add a significant part to the commercial traffic volume, especially in urban areas. The reasons, commercial passenger transport lacks behind are wide-ranging, one of the leading causes being difficulties in gathering sufficient data. In this paper, we present a microscopic approach to model commercial travel demand, including but not limited to freight traffic, based on data from a national survey and open data. We differentiate between vehicles of businesses that have a fixed daily schedule, with only small variations of their trip purposes and vehicles of businesses that can predict their daily schedules only to a certain degree. The latter have varying trip purposes and decide on a short-term base if and what sort of trip is to be pursued. Vehicles with fixed daily schedules include plumbers, electricians, care services, and delivery trucks. Due to our database, we produced a model for these vehicles exemplary for delivery by determining the number of trips for a day and assigning destinations to those trips afterward. We also take the number of private trips into account, laying the foundation of being able to incorporate the commercial transport model into a passenger transport model. We show that our model can overcome the lack of regional data. Based on generic data, the application of our approach shows promising results for the urban and regional commercial travel demand of a model region. By basing our model on generic data, we introduced an opportunity to model commercial travel demand not only in one model region but also for other urban areas in Germany and possibly in various areas in Europe, assuming that structural data is similar.

Who Uses Freeways and Who Pays for Them? Model-Based Analysis of Distribution Effects of Toll Tariff Systems in Germany

The introduction of a car toll on German freeways is a current political discussion and raises several questions, including whether particular car owner groups are financially disadvantaged by certain toll tariff systems. Analyses of this issue require a detailed knowledge of the frequency and intensity of freeway usage across the car fleet from a longitudinal perspective. Because such data were not available for Germany, this study developed an approach that extended a microscopic data set of car trips for a 1-year period by modeling the highway usage of those trips. The resulting annual toll expenses for every car in the sample were evaluated for three toll tariff systems: an annual fee, combinations of tariffs for different periods, and a tariff that depended on the vehicle miles traveled (VMT). The analyses illustrated that the toll expenses per car varied greatly between the systems. Car owners with low annual mileages were financially disadvantaged by the annual sticker toll tariff. Comparisons of toll expenses and the socioeconomic aspects of car owners and car characteristics showed similar results. Car owners with a low income and retired or unemployed persons would benefit notably from the VMT toll.

Carsharing - ein neues Verkehrssystem!

Die Bedeutung von Carsharing als Verkehrsmittel nimmt zu. Insbesondere in grosstadischen Bereichen ist Carsharing Teil der stadtischen Mobilitat, Tendenz steigend. In Deutschland ist die Anzahl an Carsharing-Nutzern und der in den Carsharing-Systemen angebotenen Fahrzeuge in den letzten Jahren stetig gestiegen und Prognosen bestatigen diesen Trend.