Studying Driving Risk Factors using Multi-Source Mobile Computing Data

Abstract

Traffic congestion can largely be attributed to the issues related with driving behavior, which may cause vehicle crash, stop-and-go traffic due to frequent lane changing behaviors, etc., and makes the driving behavior research also of significance in the realm of traffic management and demand management. The emergence and subsequent rapid advances with new information and communication technologies (ICT) now offers the capability of collecting high-fidelity and highresolution trajectory data in a cost-effective manner. In this research, we use a smartphone app to collect data for the purpose of studying driving risk factors. What's unique about the data in this research is its backend server also estimates traffic speed and volume for each link that the vehicle traverses. In order words, the data collected with build-in GPS modules in the smartphone include not only the vehicle spatial-temporal dimension location, which could be used to correlate the network geography attributes and/or real-time traffic condition, but also the detailed information about the vehicle dynamics including speed, acceleration, and deceleration, whereby a driver's control and maneuver of a vehicle can be analyzed in detail. Such type of dataset combining both user trajectory and link speed/volume information is rarely seen in prior research, permitting a unique opportunity to link critical traffic congestion factors leading to driving behavior and crash potential. In this paper, the overall research framework used in this research is presented, which mainly includes data collection, data processing, calibration and analysis methodology. A preliminary case study — including data summary statistics and correlation analysis — is also presented. The results of our study will further existing knowledge about driving exposure factors that are closely linked to crash risk, and provide the foundation for advanced forms of Usage Based Insurance.