Driving Behavior Analysis with Smartphones: Insights from a Controlled Field Study
ISBN
978-1-4503-1815-0
Type
conference paper
Date Issued
2012-12-06
Author(s)
Abstract (De)
We evaluate a mobile application that assesses driving behavior based on in-vehicle acceleration measurements and gives corresponding feedback to drivers. In the insurance business, such applications have recently gained traction as a viable alternative to the monitoring of drivers via "black boxes" installed in vehicles, which lacks interaction opportunities and is perceived as privacy intrusive by policyholders. However, pose uncertainty and other noise-inducing factors make smartphones potentially less reliable as sensor platforms. We therefore compare critical driving events generated by a smartphone with reference measurements from a vehicle-fixed IMU in a controlled field study. The study was designed to capture driver variability under real-world conditions, while minimizing the influence of external factors. We find that the mobile measurements tend to overestimate critical driving events, possibly due to deviation from the calibrated initial device pose. While weather and daytime do not appear to influence event counts, road type is a significant factor that is not considered in most current state-of-the-art implementations.
Language
German
Keywords
Measurement
Design
Reliability
Experimentation
Human Factors
Mobile Sensing
Smartphone
IMU Data
Driving Behavior
Insurance
Pattern Recognition
Social and Behavioral Sciences
HSG Classification
contribution to scientific community
Refereed
Yes
Book title
Proceedings of the 11th International Conference on Mobile and Ubiquitous Multimedia
Publisher
ACM
Publisher place
New York, USA
Event Title
11th ACM SIGMOBILE Conference on Mobile and Ubiquitous Multimedia (MUM'12)
Event Location
Ulm, Germany
Event Date
04.-06.12.2012
Subject(s)
Division(s)
Eprints ID
218703
File(s)![Thumbnail Image]()
Name
Paefgen et al. - Driver Behavior Analysis with Smartphones.pdf
Size
719.07 KB
Format
Adobe PDF
Checksum (MD5)
9da1ecd348650aec680221211e3d5525