Traffic management nowadays is a complex task where pure monitoring of the network is not sufficient anymore for network operation. To support the road authorities and traffic engineers to optimize the traffic operation models are developed to estimate traffic states and to predict future traffic states in advance to give the possibility to anticipate with the traffic and not just to react on it. Various models from traffic modeling, control, and mathematics are building a suite of tools to support traffic engineers in the optimization process. These tools are developed by experts in different areas and this thesis deals with the design and development of a framework that enables the user to combine these tools to a system with a single user interface and the flexibility of changing modules independent from the others. This allows researchers to test new approaches and algorithms in a stable environment and allows practitioners to use state of the art modules as soon as they are available without a time consuming reengineering of the existing system or changes in the system hardware. The basic input for online traffic management systems are real-time measurements, consisting of traffic conditions, traffic control messages, controller programs and weather information. These measurements come from different systems on the roadside and are stored in specific formats and locations. To cope with this variety, we developed a database structure to store the collected data in a common structure so that the handling of the data can always be performed in the same way. At a further step the database is extended so that not only the measurements, but also the network itself can be stored with all elements. Therewith, we created a single data source for traffic management systems. (Author/publisher)
Abstract