Proactive Detection of Crash Hotspots Using In-Vehicle Driving Recorder

Crash hotspot detection is important to reduce traffic crashes by allowing effective deployment of countermeasures in those locations. However, current hotspot detection methods rely mostly on crash occurrences and, therefore, countermeasures can be implemented only after a number of crashes have been occurred. To prevent crashes prior to their actual occurrences, crash precedents, also known as surrogate safety measures, are required. In this regard, driving behavior is recognized as a reliable precedent of crash occurrence because it reflects how human drivers respond to their driving environments. Therefore, the objective of this study is to develop a proactive crash hotspot detection method by evaluating relation between crash and aggressive driving behavior that was extracted from 5 taxi drivers' driving records. The spatial correlation analysis results showed that there was a distinct positive correlation between crash and aggressive behavior occurrence as evident from high correlation coefficient (which is about 0. 863.) This finding implies that aggressive driving incidents that were suggested in this research could be considered as a surrogate safety measure for proactively detecting crash hotspots.

[1]  V. Prasannakumar,et al.  Spatio-Temporal Clustering of Road Accidents: GIS Based Analysis and Assessment , 2011 .

[2]  David W. S. Wong,et al.  Statistical Analysis with ArcView GIS , 2000 .

[3]  B A Jonah,et al.  Accident risk and risk-taking behaviour among young drivers. , 1986, Accident; analysis and prevention.

[4]  Geert Wets,et al.  Traffic accident segmentation by means of latent class clustering. , 2008, Accident; analysis and prevention.

[5]  Amy Loutfi,et al.  A review of unsupervised feature learning and deep learning for time-series modeling , 2014, Pattern Recognit. Lett..

[6]  H Lum,et al.  Interactive Highway Safety Design Model: accident predictive module , 1994 .

[7]  Kitae Jang,et al.  Investigation of Attributes of Kinematic Waves Preceding Traffic Collisions , 2010 .

[8]  E Scott Geller,et al.  An instrumented vehicle assessment of problem behavior and driving style: do younger males really take more risks? , 2002, Accident; analysis and prevention.

[9]  Simon Washington,et al.  Proactive detection of high collision concentration locations on highways , 2011 .

[10]  Tessa K Anderson,et al.  Kernel density estimation and K-means clustering to profile road accident hotspots. , 2009, Accident; analysis and prevention.

[11]  Bernard W. Silverman,et al.  Density Estimation for Statistics and Data Analysis , 1987 .

[12]  J. Rodgers,et al.  Thirteen ways to look at the correlation coefficient , 1988 .

[13]  Esa Alhoniemi,et al.  Clustering of the self-organizing map , 2000, IEEE Trans. Neural Networks Learn. Syst..

[14]  Nigel Collier,et al.  Change-Point Detection in Time-Series Data by Relative Density-Ratio Estimation , 2012, Neural Networks.

[15]  이주영,et al.  (The) study of large-scale digital tachograph analysis for machine learning-based aggressive driving behavior classification = 기계학습 기반 위험운전행동 분류를 위한 대규모 차량운행기록 분석 연구 , 2016 .

[16]  Shin Hyoung Park,et al.  Spatial analysis methods for identifying hazardous locations on expressways in Korea , 2012 .