Virtual Mobility – simulation for optimum solutions

By Peter Fischer

No matter if the goal is increasing the efficiency of existing transport facilities or planning new ones – for mobility experts computer simulations have become an indispensable tool. But to make sure that the virtual representation of the road users’ interactions actually provides the right answers, they need one thing in particular besides high-performance software: a lot of experience.

Not long ago the pocket calculator was among the traffic planner's key tools. And the days of the good old deterministic planning are not quite over yet because some questions can still be satisfactorily solved by simplifying the scenario so that calculations based on mathematical formulas produce the required results. However, the more complex the road network, especially in fast growing agglomerations, the more the architects of mobility profit from the new options and resources of our digital age. Using modern computer simulations, even traffic scenarios involving a multitude of interacting factors can be reproduced so realistically that there are hardly any surprises on the way from theory to practice.

Depending on the starting situation and the goal of the investigation, two main modeling methods are used: macroscopic and microscopic simulation. The first method treats traffic as a homogeneous mass, somewhat like water flowing through a pipe. The other looks in detail at the motion of each individual vehicle-driver unit. At this microscopic level, the software calculates everything that happens on the highway, tenth of a second by tenth of a second. Who is overtaking whom? Who changes lane and when? The program takes account of the type of vehicle as well as of the temperament of individual motorists.

Apart from efficiency, the outstanding strength of computer simulation is its versatility, as a quick look at the most important applications will show. Traffic planning and optimization are among the disciplines that profit strongly from the possibility to put their concepts to the virtual test in a wide range of tasks, for example:

• Feasibility studies: Early evaluation of greenfield construction projects increases investment and planning security.
• Forecast scenarios: Stress tests of existing traffic control systems, for instance when faced with changes in traffic density, enables the early identification of weaknesses.
• Guidance systems and traffic routing: Preliminary investigations provide a precise assessment of the measure's impact on the overall system.
• Strategy management measures: Simulation allows test runs without any negative impact on real-life traffic.
• Installation of adaptive network or signal control systems: Virtual fitness tests instantly reveal any potential for optimization.
• Environment-sensitive traffic control systems: Mapping of traffic-related pollution provides an optimum basis for defining traffic guidance and control measures.

The first step of a simulation project is always the calibration and validation of the traffic flow model used.

In other words, the model must accurately represent the actual situation. As long as the simulation is based on measured data, either from permanent counters on motorways or from inductive loops in urban areas, there is a solid base. But the higher the reliance on assumptions already at this stage, the thinner the ice becomes. No matter how bewitching the animation, one must never let oneself be fooled when this is the case.

"In the wrong hands, a simulation can be an elegant way of misleading an audience," says Professor Dr. Werner Brilon, former chairman of the Scientific Commission of the German Federal Ministry for Transport, Construction and Urban Development. "Even a minor error in one of the many input parameters may, under certain circumstances, lead to completely false results. So the extent to which I believe what appears on my monitor is a matter of trust - trust in the expertise and experience of the person who uses the model in question."

In the right hands, however, simulation may at times even be more reliable than reality - at least in a sense, as Professor Brilon explains: "When optimizing signal controls - which is currently the most important application of micro-simulation - carefully planned and executed models present a very small risk of error. In fact, I would even go so far as to say that if there are discrepancies between the simulated results and those actually measured later, they generally have less to do with the quality of the simulation than with the fact that the physical engineering measures carried out do not correspond entirely to the original plan."

Biography

Peer Fischer is a Vice President for Middle East and Asia at Siemens Traffic Solutions, where he has been responsible for the Siemens TS global portfolio since 2005.

He got his Master of Information Technology and Engineering at the University of Cooperative Education in Mannheim. Fischer lives in Munich, Germany, and regularly commutes between the Siemens Headquarter in Munich and his regions.