0000000000813974
AUTHOR
A. Granà
A statistical approach for calibrating a microsimulation model for freeways
In this paper the calibration of a traffic microsimulation model based on speed-density relationships is presented. Hypothesis test was applied in the calibration process to measure the closeness between empirical data and simulation outputs and determine whether the difference between (observed and simulated) speed-density relationships was statistically significant. Statistical regressions between the variables of traffic flow were developed by using traffic data observed at the A22 Brenner Freeway, Italy. Similar relationships were obtained for a test freeway segment in uncongested conditions of traffic flow by using the Aimsun microscopic simulator; thus on field conditions were reprodu…
An Italian experience on crash modeling for roundabouts
In the last few years a considerable amount of safety models and evaluation tasks have been developed and specifically dedicated to roundabouts. Several safety performance functions (SPFs), indeed, have been implemented for roundabouts worldwide. Since SPFs are developed using crashes, traffic volume and other characteristics of a specific site (or geographical area), their direct transferability to other contexts different from those in which SPFs were calibrated is not always possible and, in any case, it must be done very carefully. A safety performance function cannot be used without a transferability evaluation for sites not included in the geographic area for which it was developed. S…
Application of a genetic algorithm in calibration of traffic microsimulation models
This chapter describes the genetic algorithm-based calibration procedure for a microscopic traffic simulation model focusing on freeways and modern roundabouts. For both case studies, the genetic algorithm tool in MATLAB® was applied in order to reach the convergence between the outputs from Aimsun microscopic simulator and the observed data. The automatic interaction with Aimsun software was implemented through an original external Python script. Results showed that the genetic algorithm-based calibration procedure gave a better match to the observed data than simple manual calibration and the efficiency of the calibration efforts resulted significantly improved. At last, the calbrated mod…