Urban traffic patterns.

Urban traffic patterns. Urban traffic information centers provide a lot of information about the traffic situation on the urban transport network. This information can be used to detect traffic patterns. These traffic patterns can be used for the development of traffic management scenario’s and for traffic forecasts. This paper describes a method to analyze traffic patterns. First of all, traffic patterns are defined. Our focus is on daily traffic profiles that are defined in two ways. According to the first definition, a daily flow profile consists of a series of (quarter hourly) traffic flows. The second definition uses peak hour flows, daily traffic flow, time interval of peak periods and ratios between periods of the day as defining variables. Clustering analysis is used to distinguish different traffic patterns. In a preliminary analysis, a limited amount of days is clustered. When hourly traffic flows are used as features, the clustering leads to distinctive patterns concerning the height and shape of the patterns. The variation within the patterns is however substantial. Probably, a clustering based on peak flows, daily flows, peak times and ratios provides better results. The construction industry holiday appears to be responsible for one of the resulting clusters. Furthermore, Fridays are in general divided in a distinct cluster from Mondays, Tuesdays and Thursdays. This information can be used to estimate daily traffic patterns in advance. (Author/publisher)