Title: Modelling the driving speed on expressway ramps
Authors: Jan Elgner - CDV - Transport Research Centre (Czech Republic) [presenting]
Veronika Rimalova - Palacky University Olomouc (Czech Republic)
Eva Fiserova - Palacky University (Czech Republic)
Jiri Ambros - CDV - Transport Research Centre (Czech Republic)
Abstract: The question whether there are any causal factors affecting the speed of vehicles in road traffic is of great interest. Identifying such causal effects encounters some difficulties, due to wide range of aspects affecting vehicle drivers, e.g. road traffic volume, road curvature and cross-section characteristics, type of vehicle, road type, weather condition, etc. The example data set was collected from GPS on-board units of floating car fleet on 6 expressway ramps in the city of Brno, Czech Republic for half a year. It contains approximately 400 unique drives on each ramp. The vehicle speed data were recorded every 0.25 seconds, representing a nearly continuous phenomenon, and therefore it is convenient to employ the functional analysis (FDA) techniques. The development of FDA arose from the need for analysing complex data structures, such as curves or surfaces, and many statistical methods have been modified for functional data. In this particular case, the speed profiles are treated as functions of distance. The main interest lies in finding patterns in driving behaviour on the expressway ramps. This can be reached by examining functions derivatives, or by finding relation between speed profile and other influential aspects, either scalar or functional.