Applying a Macroscopic Calibration to a Wiedemann Driver Behavior Threshold Study

Abstract

A speed - flow graph is used to calibrate Wiedemann 74 model parameters. A parameter calibration platform integrating VISSIM, Matlab, and ExcelVBA is built, which minimizes the differences between the measured values and the simulated values of the speed - flow graph. The optimization methods adopt the relevant knowledge in image recognition and a genetic algorithm. This platform achieves automatic parameter optimization by iteration. The proposed parameter calibration platform calibrates the driver behavior threshold based on a speed - flow graph, provides an automated technique for parameter calibration using detector data, and provides an effective method for studying driver behavior. This platform solves the low precision problem of default parameters, which is not suitable for Chinese traffic conditions. We collect traffic flow data at the south second ring road of Changsha city using a roadside laser detector. We first calibrate Wiedemann’s model using collected data and the proposed parameter calibration platform. Then, we fit the driver behavior threshold curve. Finally, we analyze driver behavior passing Changsha south second ring road.