Uncovering the Complex Causal Mechanisms of Road Traffic Collisions at Intersections Using Piecewise Structural Equation Modelling

Yıl 2024, Cilt: 2 Sayı: 2, 125 - 136, 30.12.2024

Mustafa Ekmekçi , Nima Dadashzadeh Lee Woods Renan Sinanmis Balci

Öz

Understanding the causes of traffic collisions is crucial for road designers, engineers, and policymakers to improve road safety at intersections. Design standards aim to minimize the severity and frequency of collisions. However, the factors that may affect traffic collisions are extensive. Their causal mechanisms can be complex, with feedback loops between traffic flows, visibilities, speeds, risk perception, speed limits, and other geometric characteristics of intersections. Structural Equation Modelling (SEM) is commonly used in behavioural sciences to understand complex causal paths, including travel behaviour studies. However, SEMs cannot robustly represent non-normally distributed datasets and rare count events, and little literature exists on their application to road traffic collisions. To address this limitation, this paper proposes a piecewise Structural Equation Modelling (pSEM) technique, which can handle count responses (i.e. number of collisions) to represent the complex causal relationships that lead to collisions. Application of pSEM technique is compared with conventional SEM. The Akaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC) values demonstrate that pSEM is a more robust approach to model collisions at unsignalized intersections than conventional SEM. In terms of prediction ability, AutoML is much more robust than pSEM and SEM. However, due to difficulties of interpretation for AutoML, it is not recommended for policy implications.

Anahtar Kelimeler

Road safety, Traffic collision analysis, Piecewise Structural Equation Modeling (pSEM), Intersection design, Priority three-armed intersections

Destekleyen Kurum

Çanakkale Onsekiz Mart University, The Republic of Türkiye Ministry of National Education

Teşekkür

This work is fully funded by the Republic of Türkiye Ministry of National Education, Çanakkale Onsekiz Mart University: Study Abroad Program.

Kaynakça

• [1]Department for Transport. Road Safety Data - data.gov.uk [Internet]. 2023 [cited 2024 Jan 15]. Available from:https://www.data.gov.uk/dataset/cb7ae6f0-4be6-4935-9277-47e5ce24a11f/road-safety-data
• [2]Eenink R, Reurings M, Elvik R, Cardoso J, Wichert S, Stefan C.Accident Prediction Models and Road Safety ImpactAssessment: recommendations for using these tools. Vol.506184, RiPCORD iSEREST. 2005.
• [3] Usman T, Fu L, Miranda-moreno LF. Quantifying safety benefit of winter road maintenance: Accident frequency modeling. AccidAnal Prev [Internet]. 2010;42(6):1878-87. Available from:http://dx.doi.org/10.1016/j.aap.2010.05. 008
• [4]Sarkar A, Sahoo UC, Sahoo G. Accident prediction models forurban roads. Int J Veh Saf. 2012;6(2):149-61.
• [5]Çelik AK, Oktay E. A multinomial logit analysis of risk factorsinfluencing road traffic injury severities in the Erzurum and Kars Provinces of Turkey. Accid Anal Prev. 2014 Nov 1;72:66-77.
• [6]Salifu M. Analysis of accident potential at unsignalised urbanjunctions in Ghana. 2002.
• [7]Sayed T, Rodriguez F. Accident prediction models for urbanunsignalized intersections in British Columbia. Transp Res Rec. 1999;(1665):93-9.
• [8]Muthen LK. Mplus Discussion >> Count Data Within PathModels [Internet]. 2012 [cited 2022 Apr 26]. Available from:http://www.statmodel.com/discussion/messages/11/802.html?1562474031
• [9]Kitto HJ. Accident rate at urban right-angle intersections.1980.
• [10]Bonneson JA, Mccoy PT. Estimation of Safety at Two-WayStop-Controlled Intersections on Rural Highways. 1993.
• [11]Kulmala R. Safety at rural three- and four-arm junctions.Development and application of accident prediction models. VTT Publications [Internet]. 1996 [cited 2021 Mar 25];(233).Available from: https://trid.trb.org/view/684994
• [12]Chen H, Cao L, Logan DB. Analysis of Risk Factors Affecting theSeverity of Intersection Crashes by Logistic Regression.http://dx.doi.org/101080/153895882011653841 [Internet]. 2012May [cited 2022 Aug 19];13(3):300-7. Available from:https://www.tandfonline.com/doi/abs/10.1080/15389588.2011.653841
• [13]Haleem K, Abdel-Aty M. Examining traffic crash injuryseverity at unsignalized intersections. J Safety Res. 2010Aug;41(4):347-57.
• [14] Zhou D, Gayah VV, Wood JS. Integration of machine learning and statistical models for crash frequency modeling.Transportation Letters. 2023;15(10):1408-19.
• [15]Akin D, Akbas B. A neural network (NN) model to predictintersection crashes based upon driver, vehicle and roadwaysurface characteristics. 2010; Available from: https://www. researchgate.net/publication/228367341
• [16]Polus A. Driver behaviour and accident records atunsignalized urban intersections. Accid Anal Prev. 1985 Feb1;17(1):25-32.
• [17]Caliendo C, Guida M. Microsimulation approach forpredicting crashes at unsignalized intersections using trafficconflicts. J Transp Eng. 2012;138(12):1453-67.
• [18]Mukund Pawar N, Ninad Gore;, Arkatkar S. ExaminingCrossing Conflicts by Vehicle Type at Unsignalized T-Intersections Using Accepted Gaps: A Perspective fromEmerging Countries. J Transp Eng A Syst. 2022 Jun;148(6).
• [19]Goyani J, Asce SM, Aninda;, Paul B, Gore N, Arkatkar S, et al.Investigation of Crossing Conflicts by Vehicle Type at Unsignalized T-Intersections under Varying Roadway and Traffic Conditions in India. J Transp Eng A Syst. 2021 Feb;147(2).
• [20] Paul AB, Goyani J, Arkatkar S, Joshi G. Modeling the Effect of Motorized Two-Wheelers and Autorickshaws on CrossingConflicts at Urban Unsignalized T-Intersections in India using Surrogate Safety Measures. Transportation ResearchProcedia. 2022;62:774-81.
• [21] Pirdavani A, Brijs T, Bellemans T, Pirdavani A, Brijs T, Bellemans T, et al. Evaluation of Traffic Safety at Un-signalized Intersections UsingMicrosimulation: A Utilization of Proximal Safety Indicators[Internet]. Vol. 22, Advances in Transportation Studies aninternational Journal Section A. 2010. Available from:https://www.researchgate.net/ publication/228660289
• [22]Srinivasula SR, Chepuri A, Arkatkar SS, Joshi G. Developingproximal safety indicators for assessment of un-signalizedintersection – a case study in Surat city. TransportationLetters [Internet]. 2020 May 27 [cited 2023 Feb22];12(5):303-15. Available from: https://www.tandfonline.com/action/journalInformation?journalCode=ytrl20
• [23] Orsini F, Gastaldi M, Rossi R. Conflict-Based Real-Time RoadSafety Analysis: Sensitivity to Data Collection Duration andits Implications for Model Resilience. Transp Res Rec[Internet]. 2024;2678(1):460-72. Available from: https://doi.org/10.1177/03611981231171151
• [24]Arndt O, Troutbeck R. Relationship between unsignalisedintersection geometry and accident rates - A literaturereview. Vol. 10, Road and Transport Research. 2003.
• [25]Nambuusi BB, Brijs T, Hermans E. A Review of AccidentPrediction Models for Road Intersections. Infrastructuur enruimte. 2008. 69 p.
• [26]Yannis G, Antoniou C, Papadimitriou E. Road casualties andenforcement: Distributional assumptions of seriallycorrelated count data. Traffic Inj Prev [Internet]. 2007 Sep[cited 2023 Feb 22];8(3):300-8. Available from:https://www.tandfonline.com/action/journalInformation?journalCode=gcpi20
• [27]Dinarcan GN. Count Data Regression Models 1. HacettepeÜniversitesi; 2018.
• [28]Rahman S. Development of an Accident Prediction Modelfor Intersections of Dhaka City, Bangladesh. Int J ComputAppl. 2012;47(16):10-6.
• [29]Hauer E, Ng JCN, Lovell J. Estimation of safety at signalizedintersections. Transp Res Rec. 1988;(1185):48-61.
• [30] Mahel M, Summersgill I. A comprehensive methodology forthe fitting of predictive accident models. Science (1979).1996;28(3):281-96.
• [31] Salifu M. Accident prediction models for unsignalised urban junctions in Ghana. International Association of Traffic andSafety Sciences [Internet]. 2004;28(1):68-81. Available from: http://dx.doi.org/10.1016/S0386-1112(14)60093-5
• [32]Turner S, Nicholson A, Miaou SP, Lord D. Intersectionaccident estimation: The role of intersection location andnon-collision flows. Accid Anal Prev. 2003;30(1840):31-40.
• [33]Chin HC, Quddus MA. Applying the random effect negativebinomial model to examine traffic accident occurrence atsignalized intersections. Accid Anal Prev. 2003;35(2):253-9.
• [34]Sawalha Z, Sayed T. Transferability of accident predictionmodels. Saf Sci. 2006;44(3):209-19.
• [35]Vogt A. Crash models for rural intersections: Four-lane bytwo-lane stop-controlled and two-lane by two-lanesignalized [Internet]. Vol. Jan, Transportation ResearchRecord: Journal of the Transportation Research Board. 1999[cited 2021 Jun 1]. 18-29 p. Available from:https://trid.trb.org/view/648036
• [36]Hilbe JM. Negative binomial regression [Internet]. NegativeBinomial Regression. 2007 [cited 2021 Jun 7]. 1-251 p.Available from: https://books.google.co.uk/books?hl=tr&lr=&id=0Q_ijxOEBjMC&oi=fnd&pg=PR11&dq=Poisson+regression+is+a+special+situation+of+Negative+Binomial+regression.&ots=IK2_V
• [37]Poch M, Mannering F. Negative binomial analysis ofintersection-accident frequencies. J Transp Eng.1996;122(2):105-13.
• [38]Kumara, S. S. P., and Chin HC. Modeling AccidentOccurrence at Signalized Tee-Intersections with SpecialEmphasis on Excess Zeros. 2003.
• [39]Prasetijo J, Zahidah Musa W. Modeling Zero InflatedRegression of Road Accidents at Johor Federal Road F001.2016.
• [40]Wright S. Correlation and Causation. 1921.
• [41]Wright S. The method of path coefficients. 1934.
• [42] Fornell C, Larcker DF. Evaluating Structural Equation Modelswith Unobservable Variables and Measurement Error. Vol.18, Journal of Marketing Research. JSTOR; 1981. 39 p.
• [43]Pickering D, Hall RD, Grimmer M. Accidents at rural T-junctions. Vol. P270, Research Report - Transport and RoadResearch Laboratory. 1986.
• [44]World Population Review. Portsmouth Population(Demographics, Maps, Graphs) [Internet]. 2022 [cited 2022Nov 4]. Available from: https://worldpopulationreview.com/world-cities/portsmouth-population
• [45]Bunce A. Portsmouth’s 20mph Scheme [Internet]. 2009[cited 2022 May 17]. Available from:https://www.roadsafetyknowledgecentre.org.uk/rskc-54/
• [46]CIHT. Transport in the urban environment (TUE) [Internet].1997 [cited 2022 May 19]. Available from:https://www.thenbs.com/PublicationIndex/documents/details?Pub=IHT&DocID=305156
• [47]Ekmekci M. An investigation of piecewise StructuralEquation Modelling to understand the effect on collisions ofshorter visibility designs at urban three arm junctions.[Portsmouth]: University of Portsmouth; 2023.
• [48]Shipley B. Confirmatory path analysis in a generalizedmultilevel context. Vol. 90, Ecology. 2009.
• [49] Stein CM, Morris NJ, Nock NL. Structural Equation Modeling.Methods in Molecular Biology [Internet]. 2012 [cited 2022Apr 27];850:495-512. Available from: https://link.springer.com/protocol/10.1007/978-1-61779-555-8_27
• [50]Lefcheck JS. piecewiseSEM: Piecewise structural equationmodelling in r for ecology, evolution, and systematics.Methods Ecol Evol [Internet]. 2016 May 1 [cited 2022 Aug9];7(5):573-9. Available from: https://onlinelibrary.wiley.com/doi/full/10.1111/2041-210X.12512
• [51]Shipley B. A New Inferential Test for Path Models Based onDirected Acyclic Graphs. http://dx.doi.org/101207/S15328007SEM0702_4 [Internet]. 2000 [cited 2022 Apr26];7(2):206-18. Available from: https://www.tandfonline.com/doi/abs/10.1207/S15328007SEM0702_4
• [52]Pearl J, Verma T. The logic of representing dependencies bydirected graphs. 1987.
• [53]Shipley B. The AIC model selection method applied to pathanalytic models compared using a d-separation test. Ecology[Internet]. 2013 Mar 1 [cited 2022 Apr 27];94(3):560-4.Available from: https://onlinelibrary.wiley.com/doi/full/10.1890/12-0976.1
• [54]Fisher RA. Statistical methods for research workers.Statistical methods for research workers. 1954;(10th. ed.).
• [55]Ekmekci M, Dadashzadeh N, Woods L. Assessing the impactof low-speed limit zones’ policy implications on cyclist safety: Evidence from the UK. Transp Policy (Oxf) [Internet]. 2024Jun; 152:29-39. Available from: https://linkinghub.elsevier. com/retrieve/pii/S0967070X24001173
• [56] Ekmekci M, Woods L, Dadashzadeh N. Effects of road width,radii and speeds on collisions at three-arm priorityintersections. Accid Anal Prev. 2024 May 1;199.
• [57]Chakrabarti A, Ghosh JK. AIC, BIC and Recent Advances inModel Selection. Philosophy of Statistics. 2011 Jan 1;583-605.