Entwicklung und Implementierung eines Algorithmus zur Generierung von Verkehrsmeldungen für den motorisierten Straßenverkehr, in urbanen Gebieten, basierend auf stark heterogenen Reisezeitinformationen

Kurzfassung

This Master thesis describes the development and the implementation of algorithms to generate traffic messages for the motorised traffic in urban areas. The developed algorithms have to work with strongly heterogeneous travel time information and generate from this, concise traffic messages. This applied research was accomplished at the Institute of Transport Research at the German Aerospace Centre (DLR). Since several years the researchers have worked with floating car data (FCD) and have researched the traffic situation in large cities. Taxis are used as floating cars, whereby at present approx. 5.000 vehicles deliver data from six european and one asiatic city continuously. The technology of floating car data is described in principles. Emphasis is on the algorithms, which produce finally the traffic messages. Possible representational forms and formats are pointed out. The specified earning of information is not a component of this thesis and is only principle regarded. In the context of this master thesis two different message systems (message generator, TMC generator) were developed. The Message-Generator delivers messages in the plain text for broadcast. The TMC-Generator in opposite to it, delivers messages in TMC-XML format. These are transferred to a FTP-server and coded there into TMC format. The description of the TMC coding is not the subject of this thesis. However the XML format, from which the TMC messages were generated, is described. The structure of this thesis follows in far parts the temporal succession of the activities. Only the research of the existing conventional basics for the recognition and modelling of traffic conditions and the development of own ideas for detection of congestion ran parallel. At the beginning the traffic conditions are considered theoretically. Subsequently, classical ideas and forms of the pilot research are treated. Afterwards own algorithms and ideas for detection of congestion are described. These algorithms were implemented and represented in the complex system. A validation and evaluation of the developed algorithms is finishing this thesis.