Originally designed for petrol and diesel-powered vehicles, the transport network of today is not optimised for electromobility. Very much aware of this fact, people from all over are reluctant to make the switch to e-mobility. Furthermore, investing in an extensive network of high-speed charging points would be costly. A better alternative, therefore, is intelligent use of available infrastructure and technical systems. In connection with the research project “Smart Mobility in Thuringia”, the federal state capital of Erfurt in Germany is creating an integrated traffic management system for the metropolitan region. Therefore they use PTV Optima. This model-based solution combines the tried-and-tested offline transport modelling method with real-time data and algorithms. The benefit: this development will not only benefit owners of electric vehicles, but perspectively also other road users.

“For drivers of electric vehicles, navigation which optimises journey time based on up-to-date traffic information is key”, says Frank Helbing from the Civil Engineering and Transport Agency of the state capital of Erfurt. In order to cater for them, Erfurt is developing a system which collects a whole host of information from different data sources, compiles it, and then delivers it to the driver of the electric vehicle via the “sMobiliTy-Cloud” in the form of high-quality traffic information. The state capital chosePTV Optima for the implementation of this project. “Erfurt has an advanced traffic model, which has been created and refined by the Urban Planning and Development Agency andverkehrplus GmbH using PTV Visum over years”, reports Frank Helbing. “This traffic model uses PTV Optima to form the basis for the traffic forecasts.”

Demand-based modelling is the key

The transport model created in PTV Visum shows each „typical day” such as working days or weekends fort he selected transport area. It models transport services and travel demand using demand matrices. Dynamic traffic assignment is used to calculate the time-related traffic volume and turning movements in networks based on travel demand. Then, all information is transferred from PTV Visum to PTV Optima. This is where online data comes into play. In PTV Optima, the data is used in real time in order to adjust capacity, speed and volume from PTV Visum’s base model to the current local flow and road conditions.

As PTV Optima explicitly includes the network structure, traffic flow dynamics and the route choice behaviour of road users, it also covers the traffic situation for routes without detectors (spatial distribution) and predicts the impacts of unforeseen incidents (temporal distribution) up to 60 minutes. Moreover one can assess and compare different strategic actions.

Figure-1 Integration-574x1024

“Prior to the online project, our traffic model contained detailed demand matrices for every hour of every working day”, reports Frank Helbing. “Now, together with verkehrplus GmbHwe have also added demand matrices for Saturdays and Sundays, and extended the network of measuring points.” To do so, detailed modelling was carried out, and around 1800 detectors were mapped, particularly in the areas surrounding intersections. In fact, a quarter of these have already been activated. In the following months also floating car data (FCD) will be integrated. “It is important to us that the project is built on the basis of demand”, says Frank Helbing. “We were convinced by PTV Optima’s innovative approach”

Statistical vs. simulation modelling approach

There are two main, distinct approaches to forecasting traffic situations: the statistical approach and the model-based approach. The statistical modelling approach uses interpolation, interference, data mining, “artificial intelligence” and mathematical models to match the observed time series with historical patterns. Flow and speed variables are analysed and forecasted without trying to explain and reproduce the underlying phenomena, namely vehicular interaction and driver behaviour. Statistical modelling techniques can be used for predicting traffic measures in low volatility or homogeneous traffic patterns with random variable discovery methods. However, this technique fails, if there is not enough historical data, which is often the case in unusual situations, e.g. accidents or road works. As a result, the statistical modelling approach is not appropriate for those traffic conditions. Then the operators will need more support.

Conversely, the model-based approach is based on an explicit and physical interpretation of the network, demand and traffic conditions. The underlying model which describes these interactions allows a simulation to calculate information that are either difficult or impossible to measure. An example for difficult measurements would be to observe the traffic state of the entire network. While, it is utterly impossible to measure future traffic conditions or the effects of various combinations of traffic management measures and incidents.

As a result,PTV Optima allows an effective and consistent representation and prediction of impacts such as drops in speed, spillbacks or queue pattern generated by unexpected (single event or a combination of events which never happened before. Moreover the solution enables to capture on-trip re-routing of motorists caught behind an incident in a more realistic way as well as to estimate the effect and predict the impact of signal control changes on traffic conditions. “We are currently in the midst of the development process, and at the moment we are focusing on our first step: providing essential traffic information for e-vehicle drivers to mobile telephones and tablets, which will then also be integrated into navigation”, says Frank Helbing. “But, of course, we are also toying with the idea of how to use the system for other purposes in the future. In a city that has grown over the years, we need to use the existing transport network as efficiently as possible. Powerful and dynamic traffic management is the key to this.”


Original source: PTV Group