Assessment of Expressway Traffic Safety Using Gaussian Mixture Model based on Time to Collision

Traffic safety is of great significance, especially in urban expressway where traffic volume is large and traffic conflicts are highlighted. It is thus important to develop a methodology that is able to assess traffic safety. In this paper, we first analyze the time to collision (TTC) samples from traffic videos collected from Beijing expressway with different locations, lanes, and traffic conditions. Accordingly, some basic descriptive statistics of 5 locations’ TTC samples are shown, and it is concluded that Gaussian mixture model (GMM) distribution is the best-fitted distribution to TTC samples based on K-S goodness of fit tests. Using GMM distribution, TTC samples can be divided into three categories: dangerous situations, relative safe situations, and absolute safe situations, respectively. We then proceeds to introduce a novel concept of the percentage of serious traffic conflicts as the percentage of TTC samples below a predetermined threshold value in dangerous situation. After that, assessment results of expressway traffic safety are presented using the proposed traffic safety indictor. The results imply that traffic safety on the weaving segment is lower than that on mainlines and the percentage of serious traffic conflicts on median lane is larger than that on middle lane and shoulder lane.

[1]  R. Redner,et al.  Mixture densities, maximum likelihood, and the EM algorithm , 1984 .

[2]  Bart van Arem,et al.  The effect of a special lane for intelligent vehicles on traffic flows , 1997 .

[3]  Fred L. Mannering,et al.  The statistical analysis of crash-frequency data: A review and assessment of methodological alternatives , 2010 .

[4]  C. Hydén.  THE DEVELOPMENT OF A METHOD FOR TRAFFIC SAFETY EVALUATION: THE SWEDISH TRAFFIC CONFLICTS TECHNIQUE , 1987 .

[5]  John C Hayward,et al.  NEAR-MISS DETERMINATION THROUGH USE OF A SCALE OF DANGER , 1972 .

[6]  Tae Jin Kang,et al.  Texture classification and segmentation using wavelet packet frame and Gaussian mixture model , 2007, Pattern Recognit..

[7]  A Heino,et al.  Choice of time-headway in car-following and the role of time-to-collision information in braking. , 1996, Ergonomics.

[8]  W. Eric L. Grimson,et al.  Adaptive background mixture models for real-time tracking , 1999, Proceedings. 1999 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No PR00149).

[9]  Robert A. Jacobs,et al.  Hierarchical Mixtures of Experts and the EM Algorithm , 1993, Neural Computation.

[10]  Xuesong Wang,et al.  Modeling signalized intersection safety with corridor-level spatial correlations. , 2010, Accident; analysis and prevention.

[11]  Michael I. Jordan,et al.  Convergence results for the EM approach to mixtures of experts architectures , 1995, Neural Networks.

[12]  J Maretzke,et al.  DISTANCE WARNING AND CONTROL AS A MEANS OF INCREASING ROAD SAFETY AND EASE OF OPERATION , 1992 .

[13]  Stuart R Perkins,et al.  Traffic conflict characteristics-accident potential at intersections , 1968 .

[14]  H Lum,et al.  Modeling vehicle accidents and highway geometric design relationships. , 1993, Accident; analysis and prevention.

[15]  Tom Brijs,et al.  Explaining variation in safety performance of roundabouts. , 2010, Accident; analysis and prevention.

[16]  R. B. Potts,et al.  Car-Following Theory of Steady-State Traffic Flow , 1959 .

[17]  Lena Nilsson,et al.  AN EXPERIMENTAL EVALUATION OF IN-VEHICLE COLLISION AVOIDANCE SYSTEMS , 1991 .

[18]  Heiner Bubb,et al.  Traffic Safety through Driver Assistance and Intelligence , 2011, Int. J. Comput. Intell. Syst..

[19]  Robert Graham,et al.  The Format and Presentation of Collision Warnings , 1997 .

[20]  Wuhong Wang,et al.  Incident Tree Model and Incident Tree Analysis Method for Quantified Risk Assessment: An In-depth Accident Study in Traffic Operation , 2010 .

[21]  New York Dover,et al.  ON THE CONVERGENCE PROPERTIES OF THE EM ALGORITHM , 1983 .

[22]  D. Rubin,et al.  Maximum likelihood from incomplete data via the EM - algorithm plus discussions on the paper , 1977 .

[23]  W. V. Winsum.  THE HUMAN ELEMENT IN CAR FOLLOWING MODELS , 1999 .

[24]  P. Nurmi.  Mixture Models , 2008 .

[25]  Hoong Chor Chin,et al.  Applying Bayesian hierarchical models to examine motorcycle crashes at signalized intersections. , 2010, Accident; analysis and prevention.

[26]  Katja Vogel,et al.  A comparison of headway and time to collision as safety indicators. , 2003, Accident; analysis and prevention.

[27]  Shlomo Bekhor,et al.  Risk evaluation by modeling of passing behavior on two-lane rural highways. , 2009, Accident; analysis and prevention.

[28]  Ching-Tang Hsieh,et al.  Robust Speaker Identification System Based on Wavelet Transform and Gaussian Mixture Model , 2003, J. Inf. Sci. Eng..

[29]  X. Zhou,et al.  Optimisation of Gaussian mixture model for satellite image classification , 2006 .

[30]  Frank C. Leeming,et al.  Headway on urban streets: observational data and an intervention to decrease tailgating , 2000 .

[31]  Dianhai Wang,et al.  Non-Lane-Based Car-Following Model with Visual Angle Information , 2011 .

[32]  M. Stephens.  EDF Statistics for Goodness of Fit and Some Comparisons , 1974 .

[33]  Thomas Fang Zheng,et al.  A tree-based kernel selection approach to efficient Gaussian mixture model-universal background model based speaker identification , 2006, Speech Commun..

[34]  M M Minderhoud,et al.  Extended time-to-collision measures for road traffic safety assessment. , 2001, Accident; analysis and prevention.

[35]  A. F. Smith,et al.  Statistical analysis of finite mixture distributions , 1986 .