Applying Bayesian hierarchical models to examine motorcycle crashes at signalized intersections (original) (raw)
Related papers
Accident Analysis & Prevention, 2014
The influence of intersection features on safety has been examined extensively because intersections experience a relatively large proportion of motor vehicle conflicts and crashes. Although there are distinct differences between passenger cars and large trucks-size, operating characteristics, dimensions, and weight-modeling crash counts across vehicle types is rarely addressed. This paper develops and presents a multivariate regression model of crash frequencies by collision vehicle type using crash data for urban signalized intersections in Tennessee. In addition, the performance of univariate Poisson-lognormal (UVPLN), multivariate Poisson (MVP), and multivariate Poisson-lognormal (MVPLN) regression models in establishing the relationship between crashes, traffic factors, and geometric design of roadway intersections is investigated. Bayesian methods are used to estimate the unknown parameters of these models. The evaluation results suggest that the MVPLN model possesses most of the desirable statistical properties in developing the relationships. Compared to the UVPLN and MVP models, the MVPLN model better identifies significant factors and predicts crash frequencies. The findings suggest that traffic volume, truck percentage, lighting condition, and intersection angle significantly affect intersection safety. Important differences in car, car-truck, and truck crash frequencies with respect to various risk factors were found to exist between models. The paper provides some new or more comprehensive observations that have not been covered in previous studies.
Accident Analysis & Prevention, 2009
Research suggested that motorists' right-of-way (ROW) violation in automobile-motorcycle gapacceptance accidents at priority (i.e., stop-/yield-controlled) T-intersections has been a safety concern to motorcyclists. This study examines the characteristics of automobile-motorcycle gap-acceptance accidents that occurred at such locations. British Stats19 accident injury database during 1991-2005 are examined in detail. Automobile-motorcycle gap-acceptance accidents are classified into three crash scenarios: approach-turn, angle crossing, and angle merging crashes. Mixed (random parameters) logit models are estimated to investigate the contributory factors to motorists' ROW violation in these three crash types. Crash features are also compared among gap-acceptance accidents and other crash scenarios. The methodological approach adopted allows for the individuals within the observations to have different parameter estimates as opposed to a single parameter representing all observations (i.e., accounts for unobserved heterogeneity potentially relating to roadway/environmental characteristics, and motorist behaviours). It was found that motorcycles' ROW was more likely to be violated on non-built-up roads, and in diminished light conditions, with non-uniform effects across the observations. Elderly/female motorists appeared to be over-represented in gap-acceptance crashes. Implications of the findings are discussed.
Accident Analysis & Prevention, 2017
This study aims at contributing to the literature on pedestrian and bicyclist safety by building on the conventional count regression models to explore exogenous factors affecting pedestrian and bicyclist crashes at the macroscopic level. In the traditional count models, effects of exogenous factors on non-motorist crashes were investigated directly. However, the vulnerable road users' crashes are collisions between vehicles and nonmotorists. Thus, the exogenous factors can affect the non-motorist crashes through the non-motorists and vehicle drivers. To accommodate for the potentially different impact of exogenous factors we convert the non-motorist crash counts as the product of total crash counts and proportion of non-motorist crashes and formulate a joint model of the negative binomial (NB) model and the logit model to deal with the two parts, respectively. The formulated joint model is estimated using non-motorist crash data based on the Traffic Analysis Districts (TADs) in Florida. Meanwhile, the traditional NB model is also estimated and compared with the joint model. The result indicates that the joint model provides better data fit and can identify more significant variables. Subsequently, a novel joint screening method is suggested based on the proposed model to identify hot zones for non-motorist crashes. The hot zones of non-motorist crashes are identified and divided into three types: hot zones with more dangerous driving environment only, hot zones with more hazardous walking and cycling conditions only, and hot zones with both. It is expected that the joint model and screening method can help decision makers, transportation officials, and community planners to make more efficient treatments to proactively improve pedestrian and bicyclist safety.
Crash risk models for a motorcycle-dominated traffic environment
2016
This paper presents a methodology to estimate the potentials of rear-end and sideswipe crashes for motorcycles moving in a motorcycle-dominated traffic environment on urban roads and examines their integration in the International Road Assessment Programme (iRAP) star rating system. The crash risk models developed are based on discrete choice models and traffic conflict techniques. The proposed methodology was validated using data collected on road segments from the city of Danang in Vietnam. The models’ field validation shows that the proposed methodology produces a good estimate of rear-end and sideswipe crash risk for motorcyclists and the enhanced iRAP star rating methodology produces most satisfactory results. It was found that risk factors such as front distance, longitudinal gap, lateral gap, lateral clearance, speed difference, and operating speed have a significant contribution to motorcycle crash risk and therefore they should be considered in the selection of appropriate ...
Motorcycle Fatalities Revisited: A Classical and Bayesian Analysis
2016
This paper examines the determinants of motorcycle fatality rates using panel data and classical and Bayesian statistical methods. It focuses on five variables in particular: universal helmet laws, partial helmet laws, cell phone use, suicidal propensities, and beer consumption. Universal helmet laws are found to be favored over partial helmet laws to reduce motorcycle fatality rates while cell phone use is found to be a significant contributor to motorcycle fatalities as is alcohol consumption. Suicidal propensities are also shown to contribute to these accidents.
International Journal of Environmental Research and Public Health
Motorcycle accidents can impede sustainable development due to the high fatality rate associated with motorcycle riders, particularly in developing countries. Although there has been extensive research conducted on motorcycle accidents on highways, there is a limited understanding of the factors contributing to accidents involving the most commonly used motorcycles on local roads. This study aimed to identify the root causes of fatal motorcycle accidents on local roads. The contributing factors consist of four groups: rider characteristics, maneuvers prior to the crash, temporal and environmental characteristics, and road characteristics. The study employed random parameters logit models with unobserved heterogeneity in means and variances while also incorporating the temporal instability principle. The results revealed that the data related to motorcycle accidents on local roads between 2018 and 2020 exhibited temporal variation. Numerous variables were discovered to influence the ...
Modeling Traffic Accidents at Signalized Intersections in the City of Norfolk, Va
This study was an attempt to apply a proactive approach using traffic pattern and signalized intersection characteristics to predict accident rates at signalized intersections in a city's arterial network. An earlier analysis of accident data at selected intersections within the City of Norfolk indicated that in addition to traffic volume, other controllable factors contributed to traffic accidents at specific intersections. These factors included area topography, lane patterns, type of road signs, turning lanes, etc. It is also known that administrative factors such as signal types, signal polices, road closures, etc., and maintenance factors such as road conditions, condition of the signals, condition of road signs, etc. also impact road accidents. The objective of this study was to relate these variables to accident rate and delineate variables that are statistically more significant for accident rate. Data on several topographical variables was collected in the City of Norfolk. These variables included number of lanes, turn lanes, pedestrian crossing, restricted lanes, etc. A linear regression model was used to establish relationship between these variables and the accident rate. The resulting regression model explained 60% of the variability. It also showed that four topographical variables are more important than other variables. These variables include number of lanes, number of turn lanes, presence of median and presence of permanent hazard like railway crossing. However, validation of model showed higher than expected variation. The model developed, in this study, overestimates the accident rate by 33%, thus, limiting its practical application.
PREDICTION UNDER BAYESIAN APPROACH OF CAR ACCIDENTS IN URBAN INTERSECTIONS
The increase in the number of automobiles in circulation has brought by direct consequence an increase of the interactions among different factors that contribute to the exposure of crashes, particularly in the urban intersections. For example, according to data of the city council of Toluca, in the State of Mexico, 80% of these accidents happens in intersections of urban routes (Hinojosa, 2003). This motivates a fundamental interest in the study of this phenomenon, from where possible policies or instruments could be obtained to reduce this class of incidents. In this communication three final distributions of Bayes Rule are shown (Gregory, 2005) that predict the probability of occurrence of accidents in urban intersections of the Metropolitan Zone of the City of Toluca. The first one is a specification of a Translated Poisson Mixture Model, which presents a weighed average of occurrence rates of observed car accidents. The second one is the One-variable Poisson-Gamma Model, which introduces an effect of random type to the error term of the flow variables, which is an independent variable. Finally, the third one is a Two-variable Possion-Gamma Model, which in addition to the randomness mentioned previously, considers the relation between the frequency of accidents and the vehicular flow. From statistical data of a year of reference, the parameters of these three models are calibrated, with which the final distribution is considered that predicts the occurrence of car accidents in intersections for various temporary horizons. In order to determine which of these three proposed models is more precise, observed data against estimated values by these models are compared, using the correlation coefficient and the GEH as measuring tools for carrying out this comparison.