Analysis of hourly highway traffic counts and models with derived data

Traffic characterization is based on continuity of traffic counts as well as analysis and evaluation. By examination of daily traffic on highway sections on month, season and year bases, determining traffic characterization was aimed. Among weight-in-motion (WIM) stations in Turkey, data is taken from station 2, located on Ankara-Sivrihisar highway section having the most intense data. Data are collected in 14 vehicle categories. The details as week number, weekday and season to provide studying in different bases were added. In case of lack hourly data, no traffic data for the day concerned is assumed. Analyses were carried on the bases of day, week, weekday, weekend, month, and season. Both standardized or unstandardized data are used to distinguish model coefficients for variable units or magnitudes. Modeling were done for three approaches as parabolic, trigo/hyperbolic and multi-variate integrated models. By examination of model performances, models are prioritized due to their successivity in reflecting observations. In parabolic and trigo/hyperbolic models, no relation among years could be obtained although there were relations within or between years. In parabolic models, between 0 and 11 third, between 12 and 18 second and between 19 and 22 first degree parabola had best fits however no relationship was held for between 22 and 23. In trigo/hyperbolic models, all combinations of different hour-intervals were the function of the combinations vii of ?cos?, ?cosh? and ?tanh?. In multi-variate integrated modelling, traffic characterization on ratios to daily traffic was independent to date or years.