Latest revision as of 04:38, 2 February 2021

Abstract

This study explored the use of two types of advanced driver assistance systems (ADAS) as tools for observing driving behavior. The first was a kinematic-based ADAS that uses speed and acceleration data to detect driving events such as hard braking, speeding and sharp turning. The second was a vision-based ADAS that uses video data to provide lane departure warnings (LDW), headway warnings (HW) and forward collision warnings (FCW). Data was collected for more than 4,500 trips and 2,200 driving hours during a period of 70 days. The sample consisted of 10 drivers that used both types of ADAS simultaneously. The information collected also included more than 17,000 records of various types of driving events. First, the events rates were estimated by the Poisson and the Poisson-lognormal models. Then, Pearson correlation and factor analysis were implemented to study the relationships among the events and to evaluate whether different types of events converged to describe the same behaviors. Significant correlations were observed between the braking and turning kinematic-based events and the FCW vision-based event, which converged under the same factor. High rates of these events may indicate that the person is driving in an urban style. The LDW, HW and speeding events converged to the second factor, which is more relevant in inter-urban areas. These findings, although based on a small-scale study, point to a potential for the use of commercial ADAS for driving behavior analysis. The integration of kinematic-based and vision-based information can provide deeper understanding of the measured behavior.

Original document

The different versions of the original document can be found in:

• http://dx.doi.org/10.17077/drivingassessment.1441

• https://ir.uiowa.edu/cgi/viewcontent.cgi?article=1441&context=drivingassessment

• https://ir.uiowa.edu/cgi/viewcontent.cgi?article=1441&context=drivingassessment,

https://ir.uiowa.edu/drivingassessment/2011/papers/73,

https://trid.trb.org/view/1112809,

https://core.ac.uk/display/129643529,

https://academic.microsoft.com/#/detail/771280564