For the development and effective use of ‘urban traffic control’ (including traffic information), both traffic data and insight into the urban traffic system are indispensable. To decide on where and when measures have to be taken, information is needed about traffic flows and traffic system performance at different points in time and under various circumstances. Furthermore, information on the origins and destinations and on the characteristics of the traffic enables more effective measures to be taken. Traffic data are collected at signalized intersections, but often these are not available at a central level. Furthermore, the insight into the urban traffic system is still limited. Currently initiatives are taken to centrally collect and prcess urban traffic data. One of these initiatives is the ViaContent system in the city of Almelo (www.almelo.nl). In ViaContent (www.viacontent.nl) data from loop detectors is collected in such a way that a database is constructed with data on flows and speeds for the entire urban traffic system. In this Ph.D. research, traffic data from ViaContent are used to analyse the temporal and spatial variability of urban traffic flows. Most existing research on temporal variations in traffic flows focuses on the influence of known factors (day of the week, season, Holidays, weather). Another way to take variations between days into account is by defining different types of days on the basis of measured traffic flows by means of cluster analysis. In this research cluster analysis is used for the determination and analysis of urban daily flow patterns. Firstly, it is investigated what typical daily traffic patterns can be distinguished. Secondly, it is examined what temporal and circumstantial characteristics are on the basis of these traffic patters. Finally, the variation within and between the resulting traffic patterns is analyzed to determine the applicability of the resultant traffic patterns for traffic management and traffic forecasting. The resulting traffic patterns from different locations are compared to obtain more insight into spatial variations in urban traffic flow patterns. Moreover, by means of combining data from multiple locations, temporal variations are analysed on network level as well. The resulting traffic patterns can be used as a basis for traffic management scenarios, for imputation of missing and erroneous data, as a basis for the estimation of dynamic OD-matrices, and for traffic forecasting. Moreover, information on the characteristics on the basis of the typical traffic patterns will results in a better understanding of urban traffic in general. (Author/publisher)
Samenvatting