Identifying Congestion Patterns in Urban Road Networks using Floating Car Data

Investigation of congestion patterns in urban road networks based on floating car data (FCD) has shown that noisy and partially incomplete travel speed data from floating cars can be transformed into a consistent and more precise congestion detector value. This value correlates well with measured travel times and observed congestion cases. The method computes local congestion scores, which are averaged and smoothed taking into account information of the actual vehicle paths. The method is able to deal with noisy and missing data. It has been evaluated with measured travel times and automatic congestion detection with video data from two independent systems on a major road in the city of Vienna.