Abstract:

In order to describe drivers’ behavior when driver takes an emergency lane-change action, an openloop
model was proposed based on the naturalistic driving data which was collected by the program named
China pilot Field Operation Test. Almost 228 lane-change use cases were extracted from the naturalistic driving
data. Steering wheel angle rate combined with steering wheel angle were used as filters to classify the different
emergency degrees of lane-change. The duration of drivers’ emergency lane-change was studied based
on a liner dependency between the maximum values of both steering wheel angle and steering wheel angle
rate. Parameters were fitted by the 50th percentile of the naturalistic driving data collected from lane-change
scenarios .The model can be applied in research of Chinese drivers’ behavior and its values have been proved
by both significant test and correlation test. The results of tests show that coherence of the outputs from real
driver and the model is good, so the model can well describe Chinese drivers’ emergency lane-change behavior.

Key words: automobile active safety, naturalist driving behavior, emergency lane-change, parameter fitting, driver model