GIS-based method for detecting high-crash-risk road segments using network kernel density estimation

Identifying high-crash-risk road segments provides safety specialists with an insight to better understanding of crash patterns and enhancing road safety management. The common hotspot identification methods are not robust enough to visualize the underlying shape of crash patterns since they neglect the spatial properties of crash data. Spatial traffic accidents have the tendency to be dependent, a phenomenon known as spatial autocorrelation. Values over distance are more or less similar than expected for randomly associated observations. Modeling the spatial variations can properly be explained in terms of first- and second-order properties. The first-order properties, describe the way of varying the expected value of point pattern in space which can be due to changes in the substantial properties of the local environment, while second-order effects describe the interactive effects of events explaining on how the events are interacted. Considering the discrete nature of crash data and the limited access to exact locations where crashes occur, it is likely that a continuous surface drawn from discrete points will better reflect crash density, present a more realistic picture of crash distribution. Network kernel density estimation (NKDE) is a nonparametric approach for events distributed over one-dimensional space which facilitates estimating the density at any location in the study region not just at the location where the event occurs. NKDE for road safety applications enables the extraction and visualization of crash density along roadways The application of suggested method was illustrated for Arak-Khomein rural road in Markazi province, Iran and the stability of hazardous segments by examining the resulted network estimated density during the three years of study (2006–2008) was investigated. The result of this paper helps the traffic engineers and safety specialists to determine the segments which demand more safety attentions from both transportation authorities and drivers and request assigning the resources such as budget and time.

[1]  William R. Black,et al.  Network Autocorrelation in Transport Network and Flow Systems , 2010 .

[2]  Atsuyuki Okabe,et al.  The SANET Toolbox: New Methods for Network Spatial Analysis , 2006, Trans. GIS.

[3]  D. W. Scott,et al.  Multivariate Density Estimation, Theory, Practice and Visualization , 1992 .

[4]  Jean-Claude Thill,et al.  Comparison of planar and network K-functions in traffic accident analysis , 2004 .

[5]  Ikuho Yamada,et al.  An Empirical Comparison of Edge Effect Correction Methods Applied to K -function Analysis , 2003 .

[6]  Trevor C. Bailey,et al.  Interactive Spatial Data Analysis , 1995 .

[7]  P. J. Green,et al.  Density Estimation for Statistics and Data Analysis , 1987 .

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

[9]  Alfonso Montella,et al.  A comparative analysis of hotspot identification methods. , 2010, Accident; analysis and prevention.

[10]  Joni A. Downs,et al.  Network analysis of animal space -use patterns , 2008 .

[11]  Daniel A. Griffith,et al.  Spatial-Filtering-Based Contributions to a Critique of Geographically Weighted Regression (GWR) , 2008 .

[12]  S. Erdogan.  Explorative spatial analysis of traffic accident statistics and road mortality among the provinces of Turkey. , 2009, Journal of safety research.

[13]  David W. Scott,et al.  Multivariate Density Estimation: Theory, Practice, and Visualization , 1992, Wiley Series in Probability and Statistics.

[14]  Wen Cheng,et al.  Experimental evaluation of hotspot identification methods. , 2005, Accident; analysis and prevention.

[15]  Li Zhu,et al.  A GIS-based Bayesian approach for analyzing spatial–temporal patterns of intra-city motor vehicle crashes , 2007 .

[16]  L. Anselin.  Local Indicators of Spatial Association—LISA , 2010 .

[17]  Marc Hallin,et al.  Density estimation for spatial linear processes , 2001 .

[18]  H. Miller,et al.  Representation and Spatial Analysis in Geographic Information Systems , 2003 .

[19]  Robert Haining,et al.  Spatial Data Analysis: Theory and Practice , 2003 .

[20]  OkabeAtsuyuki,et al.  A kernel density estimation method for networks, its computational method and a GIS-based tool , 2009 .

[21]  David O'Sullivan,et al.  Geographic Information Analysis , 2002 .

[22]  Lanh Tat Tran,et al.  Kernel density estimation for linear processes , 1992 .

[23]  Ibrahim Yilmaz,et al.  Geographical information systems aided traffic accident analysis system case study: city of Afyonkarahisar. , 2008, Accident; analysis and prevention.

[24]  Mohamed Abdel-Aty,et al.  Temporal and spatial analyses of rear-end crashes at signalized intersections. , 2006, Accident; analysis and prevention.

[25]  Michel Mouchart,et al.  The local spatial autocorrelation and the kernel method for identifying black zones. A comparative approach. , 2003, Accident; analysis and prevention.

[26]  Shashi S Nambisan,et al.  New methods to identify and rank high pedestrian crash zones: an illustration. , 2007, Accident; analysis and prevention.

[27]  V. Latora,et al.  Street Centrality and Densities of Retail and Services in Bologna, Italy , 2009 .

[28]  Jun Yan,et al.  Kernel Density Estimation of traffic accidents in a network space , 2008, Comput. Environ. Urban Syst..

[29]  Geert Wets,et al.  Understanding spatial concentrations of road accidents using frequent item sets. , 2005, Accident; analysis and prevention.

[30]  Giuseppe Borruso,et al.  Network Density Estimation: Analysis of Point Patterns over a Network , 2005, ICCSA.

[31]  N. Levine,et al.  Spatial analysis of Honolulu motor vehicle crashes: II. Zonal generators. , 1995, Accident; analysis and prevention.

[32]  P. Moran.  The Interpretation of Statistical Maps , 1948 .

[33]  C. Lloyd.  Local Models for Spatial Analysis , 2006 .