A crash-prediction model for road tunnels

Considerable research has been carried out into open roads to establish relationships between crashes and traffic flow, geometry of infrastructure and environmental factors, whereas crash-prediction models for road tunnels, have rarely been investigated. In addition different results have been sometimes obtained regarding the effects of traffic and geometry on crashes in road tunnels. However, most research has focused on tunnels where traffic and geometric conditions, as well as driving behaviour, differ from those in Italy. Thus, in this paper crash prediction-models that had not yet been proposed for Italian road tunnels have been developed. For the purpose, a 4-year monitoring period extending from 2006 to 2009 was considered. The tunnels investigated are single-tube ones with unidirectional traffic. The Bivariate Negative Binomial regression model, jointly applied to non-severe crashes (accidents involving material-damage only) and severe crashes (fatal and injury accidents only), was used to model the frequency of accident occurrence. The year effect on severe crashes was also analyzed by the Random Effects Binomial regression model and the Negative Multinomial regression model. Regression parameters were estimated by the Maximum Likelihood Method. The Cumulative Residual Method was used to test the adequacy of the regression model through the range of annual average daily traffic per lane. The candidate set of variables was: tunnel length (L), annual average daily traffic per lane (AADT_L), percentage of trucks (%Tr), number of lanes (N_L), and the presence of a sidewalk. Both for non-severe crashes and severe crashes, prediction-models showed that significant variables are: L, AADT_L, %Tr, and N_L. A significant year effect consisting in a systematic reduction of severe crashes over time was also detected. The analysis developed in this paper appears to be useful for many applications such as the estimation of accident reductions due to improvement in existing tunnels and/or to modifications of traffic control systems, as well as for the prediction of accidents when different tunnel design options are compared.     

Effects of road geometry and traffic volumes on rural roadway accident rates

This paper revisits the question of the relationship between rural road geometric characteristics, accident rates and their prediction, using a rigorous non-parametric statistical methodology known as hierarchical tree-based regression. The goal of this paper is twofold: first, it develops a methodology that quantitatively assesses the effects of various highway geometric characteristics on accident rates and, second, it provides a straightforward, yet fundamentally and mathematically sound way of predicting accident rates on rural roads. The results show that although the importance of isolated variables differs between two-lane and multilane roads, 'geometric design' variables and pavement condition' variables are the two most important factors affecting accident rates. Further, the methodology used in this paper allows for the explicit prediction of accident rates for given highway sections, as soon as the profile of a road section is given.     

Modeling the effects of AADT on predicting multiple-vehicle crashes at urban and suburban signalized intersections

Annual Average Daily Traffic (AADT) is often considered as a main covariate for predicting crash frequencies at urban and suburban intersections. A linear functional form is typically assumed for the Safety Performance Function (SPF) to describe the relationship between the natural logarithm of expected crash frequency and covariates derived from AADTs. Such a linearity assumption has been questioned by many researchers. This study applies Generalized Additive Models (GAMs) and Piecewise Linear Negative Binomial (PLNB) regression models to fit intersection crash data. Various covariates derived from minor-and major-approach AADTs are considered. Three different dependent variables are modeled, which are total multiple-vehicle crashes, rear-end crashes, and angle crashes. The modeling results suggest that a nonlinear functional form may be more appropriate. Also, the results show that it is important to take into consideration the joint safety effects of multiple covariates. Additionally, it is found that the ratio of minor to major-approach AADT has a varying impact on intersection safety and deserves further investigations.     

Modeling traffic accident occurrence and involvement

The Negative Binomial modeling technique was used to model the frequency of accident occurrence and involvement. Accident data over a period of 3 years, accounting for 1,606 accidents on a principal arterial in Central Florida, were used to estimate the model. The model illustrated the significance of the Annual Average Daily Traffic (AADT), degree of horizontal curvature, lane, shoulder and median widths, urban/rural, and the section's length, on the frequency of accident occurrence. Several Negative Binomial models of the frequency of accident involvement were also developed to account for the demographic characteristics of the driver (age and gender). The results showed that heavy traffic volume, speeding, narrow lane width, larger number of lanes, urban roadway sections, narrow shoulder width and reduced median width increase the likelihood for accident involvement. Subsequent elasticity computations identified the relative importance of the variables included in the models. Female drivers experience more accidents than male drivers in heavy traffic volume, reduced median width, narrow lane width, and larger number of lanes. Male drivers have greater tendency to be involved in traffic accidents while speeding. The models also indicated that young and older drivers experience more accidents than middle aged drivers in heavy traffic volume, and reduced shoulder and median widths. Younger drivers have a greater tendency of being involved in accidents on roadway curves and while speeding.     

Quasi-induced exposure: The choice of exposure metrics

The quasi-induced exposure method is widely used to estimate exposure and risks of different groups of drivers and vehicles. Essentially, this method assumes that non-at-fault or passive parties in two-vehicle collisions represent a random sample of the populations on the road. Most previous works have used the whole sample of collisions to estimate exposure.There has been some concern about possible biases in quasi-induced estimates. In this paper, we argue that (1) biases are mainly due to differences in accident avoidance abilities, speeds and injury risks, and (2) because the influence of these three factors on the probability of being non-at-fault is not the same for every crash type, differences may arise among non-at-fault populations, in which case some crash types would provide a more accurate estimate of exposure than others.We explore the direction of biases due to speed, accident avoidance ability and injury risk in four accident types: accidents between vehicles travelling on different lanes in two-way, two-lane undivided roads; accidents between vehicles travelling on different lanes on multilane roads; intersection accidents; and accidents between vehicles travelling on the same lane. Our analysis shows that more research would be needed concerning the effect of speed on head-on crashes on undivided roads, and crashes on multilane roads.     

Predicting motor vehicle crashes using Support Vector Machine models

Crash prediction models have been very popular in highway safety analyses. However, in highway safety research, the prediction of outcomes is seldom, if ever, the only research objective when estimating crash prediction models. Only very few existing methods can be used to efficiently predict motor vehicle crashes. Thus, there is a need to examine new methods for better predicting motor vehicle crashes. The objective of this study is to evaluate the application of Support Vector Machine (SVM) models for predicting motor vehicle crashes. SVM models, which are based on the statistical learning theory, are a new class of models that can be used for predicting values. To accomplish the objective of this study, Negative Binomial (NB) regression and SVM models were developed and compared using data collected on rural frontage roads in Texas. Several models were estimated using different sample sizes. The study shows that SVM models predict crash data more effectively and accurately than traditional NB models. In addition, SVM models do not over-fit the data and offer similar, if not better, performance than Back-Propagation Neural Network (BPNN) models documented in previous research. Given this characteristic and the fact that SVM models are faster to implement than BPNN models, it is suggested to use these models if the sole purpose of the study consists of predicting motor vehicle crashes.     

A novel approach for analyzing severe crash patterns on multilane highways

This study presents a novel approach for analysis of patterns in severe crashes that occur on mid-block segments of multilane highways with partially limited access. A within stratum matched crash vs. non-crash classification approach is adopted towards that end. Under this approach crashes serve as units of analysis and it does not require aggregation of crash data over arterial segments of arbitrary lengths. Also, the proposed approach does not use information on non-severe crashes and hence is not affected by under-reporting of the minor crashes. Random samples of time, day of week, and location (i.e., milepost) combinations were collected for multilane arterials in the state of Florida and matched with severe crashes from the corresponding corridor to form matched strata consisting of severe crash and non-crash cases. For these cases, geometric design/roadside and traffic characteristics were derived based on the corresponding milepost locations. Four groups of crashes, severe rear-end, lane-change related, pedestrian, and single-vehicle/off-road crashes, on multilane arterials segments were compared separately to the non-crash cases. Severe lane-change related crashes may primarily be attributed to exposure while single-vehicle crashes and pedestrian crashes have no significant relationship with the ADT (Average Daily Traffic). For severe rear-end crashes speed limit, ADT, K-factor, time of day/day of week, median type, pavement condition, and presence of horizontal curvature were significant factors. The proposed approach uses general roadway characteristics as independent variables rather than event-specific information (i.e., crash characteristics such as driver/vehicle details); it has the potential to fit within a safety evaluation framework for arterial segments.     

Linking roadway crashes and tire–pavement friction: a case study

Tire–pavement friction is a factor that can affect the rate of vehicle crashes. Several studies have suggested that reduced friction during wet weather conditions, due to water on the pavement surface reducing the contact area between the tire and the pavement, increases vehicle crashes. This study evaluates the effect of friction on both wet- and dry-condition crashes. The data for the study were provided by the New Jersey Department of Transportation. Regression analysis was performed to verify the effect of friction on the rate of wet- and dry-condition vehicle crashes for various types of urban roads. It was found that friction is not only associated with the rate of wet-condition vehicle crashes, but it also impacts the rate of dry-condition vehicle crashes. The analysis also suggested that the developed regression models could be used to define the friction demand for different road categories.     

Modeling left-turn crash occurrence at signalized intersections by conflicting patterns

In order to better understand the underlying crash mechanisms, left-turn crashes occurring at 197 four-legged signalized intersections over 6 years were classified into nine patterns based on vehicle maneuvers and then were assigned to intersection approaches. Crash frequency of each pattern was modeled at the approach level by mainly using Generalized Estimating Equations (GEE) with the Negative Binomial as the link function to account for the correlation among the crash data. GEE with a binomial logit link function was also applied for patterns with fewer crashes. The Cumulative Residuals test shows that, for correlated left-turn crashes, GEE models usually outperformed basic Negative Binomial models. The estimation results show that there are obvious differences in the factors that cause the occurrence of different left-turn collision patterns. For example, for each pattern, the traffic flows to which the colliding vehicles belong are identified to be significant. The width of the crossing distance (represented by the number of through lanes on the opposing approach of the left-turning traffic) is associated with more left-turn traffic colliding with opposing through traffic (Pattern 5), but with less left-turning traffic colliding with near-side crossing through traffic (Pattern 8). The safety effectiveness of the left-turning signal is not consistent for different crash patterns; "protected" phasing is correlated with fewer Pattern 5 crashes, but with more Pattern 8 crashes. The study indicates that in order to develop efficient countermeasures for left-turn crashes and improve safety at signalized intersections, left-turn crashes should be considered in different patterns.     

Segment characteristics and severity of head-on crashes on two-lane rural highways in Maine

More than two out of three of all fatal crashes in Maine occur on rural, two-lane collector or arterial roads. Head-on crashes on these roads account for less than 5% of the crashes, but they are responsible for almost half of all fatalities. Data analyzed in this study was provided by Maine Department of Transportation and covers all head-on crashes for 2000-2002 during which period there were 3,136 head-on crashes reported. Out of these, 127 were fatal crashes and 235 produced incapacitating but not fatal injuries. These two categories made up over 75% of the crash cost. A clear majority of head-on crashes on two-lane, rural roads in Maine were caused by drivers making errors or misjudging situations. Illegal/unsafe speed was a factor in 32% of the crashes while driver inattention/distraction was a primary factor in 28%. Fatigue was responsible for around one in 40 crashes and one in 12 fatal crashes. Alcohol or drugs was a factor in one in 12 crashes and one in nine fatal head-on crashes. Less than 8% of fatalities involved someone overtaking another vehicle, and only around 14% involved a driver intentionally crossing the centerline. Two in three fatal head-on crashes occurred on straight segments and 67% of these happened on dry pavement. There is a clear trend towards higher speed limits leading to a higher percentage of crashes becoming fatal or having incapacitating injuries. There is also a clear trend - if one keeps speeds constant and AADT within a certain range - that wider shoulders give higher crash severities. Also, for higher-speed roads, more travel lanes (than two) increase crash severity. In summary, there seems to be two major reasons why people get across the centerline and have head-on collisions: (a) people are going too fast for the roadway conditions; or (b) people are inattentive and get across the centerline more or less without noticing it. The latter category of crashes could probably be reduced if centerline rumble-strips were installed. More or less all head-on collisions could be eliminated if median barriers were installed. In-vehicle technology could also be used to significantly reduce the incidence of lane departures. Furthermore, today's speed limits should be better enforced since a high percentage of serious crashes involve illegal speeding. This should be combined with lowered speed limits for targeted high-crash segments.     

Analyzing angle crashes at unsignalized intersections using machine learning techniques

A recently developed machine learning technique, multivariate adaptive regression splines (MARS), is introduced in this study to predict vehicles’ angle crashes. MARS has a promising prediction power, and does not suffer from interpretation complexity. Negative Binomial (NB) and MARS models were fitted and compared using extensive data collected on unsignalized intersections in Florida. Two models were estimated for angle crash frequency at 3- and 4-legged unsignalized intersections. Treating crash frequency as a continuous response variable for fitting a MARS model was also examined by considering the natural logarithm of the crash frequency. Finally, combining MARS with another machine learning technique (random forest) was explored and discussed. The fitted NB angle crash models showed several significant factors that contribute to angle crash occurrence at unsignalized intersections such as, traffic volume on the major road, the upstream distance to the nearest signalized intersection, the distance between successive unsignalized intersections, median type on the major approach, percentage of trucks on the major approach, size of the intersection and the geographic location within the state. Based on the mean square prediction error (MSPE) assessment criterion, MARS outperformed the corresponding NB models. Also, using MARS for predicting continuous response variables yielded more favorable results than predicting discrete response variables. The generated MARS models showed the most promising results after screening the covariates using random forest. Based on the results of this study, MARS is recommended as an efficient technique for predicting crashes at unsignalized intersections (angle crashes in this study).     

Differences in causality factors for single and multi-vehicle crashes on two-lane roads

Past research has found a non-linear relationship between traffic intensity or level of service (LOS) and highway crash rates. This paper investigates this relationship further by including the effects of site characteristics and estimating Poisson regression models for predicting single and multi-vehicle crashes separately. Analysis focuses on rural two-lane highways, with hourly LOS, traffic composition, and highway geometric characteristics as independent variables. The resulting models for single and multi-vehicle crashes have different explanatory variables. Single-vehicle crash rates decrease with increasing traffic intensity (lower LOS), shoulder width and sight distance. Multi-vehicle crash rates increase with the number of signals, the daily single-unit truck percentage, and the shoulder width, and decreased on principal arterials compared to other roadway classes. LOS does not significantly explain variation in the number of multi-vehicle crashes. Ongoing research by the authors is aimed at identifying other site factors, such as driveway density and intersection LOS, that can better explain the differing effects reported here and predict crash rates of both types better.     

Explaining the road accident risk: Weather effects

This research aims to highlight the link between weather conditions and road accident risk at an aggregate level and on a monthly basis, in order to improve road safety monitoring at a national level. It is based on some case studies carried out in Work Package 7 on “Data analysis and synthesis” of the EU-FP6 project “SafetyNet – Building the European Road Safety Observatory”, which illustrate the use of weather variables for analysing changes in the number of road injury accidents. Time series analysis models with explanatory variables that measure the weather quantitatively were used and applied to aggregate datasets of injury accidents for France, the Netherlands and the Athens region, over periods of more than 20 years. The main results reveal significant correlations on a monthly basis between weather variables and the aggregate number of injury accidents, but the magnitude and even the sign of these correlations vary according to the type of road (motorways, rural roads or urban roads). Moreover, in the case of the interurban network in France, it appears that the rainfall effect is mainly direct on motorways – exposure being unchanged, and partly indirect on main roads – as a result of changes in exposure. Additional results obtained on a daily basis for the Athens region indicate that capturing the within-the-month variability of the weather variables and including it in a monthly model highlights the effects of extreme weather. Such findings are consistent with previous results obtained for France using a similar approach, with the exception of the negative correlation between precipitation and the number of injury accidents found for the Athens region, which is further investigated. The outlook for the approach and its added value are discussed in the conclusion.     

Evaluating safety effectiveness of surface treatment at signalised intersections: a before and after study

In a metropolitan region of Melbourne, Australia, 136 signalised intersections were identified to have been resurfaced with asphalt over the period 2005–2010. In this study, the safety effectiveness of surface treatment was evaluated using Empirical Bayes (EB) approach to account for regression to the mean bias and traffic volume change through using safety performance function (SPF). Safety effects were estimated for total casualty, high severity (fatality and serious injury) and other injury crashes. For conducting EB method a reference group was selected with similar traffic volumes and site characteristics to the treated sites. Negative Binomial regression was applied to develop SPFs that were used to predict the expected number of crashes at the treated sites. The results of EB approach revealed that the treatment effect was found to be significant at 95% confidence level for all crash severity levels. The evaluation results also showed that total casualty crashes were reduced by 21.3% with a standard error of 3.13% and high severity (fatality and serious injury) crashes were reduced by 15.3% with a standard error of 5.56%. Pavement surface treatment was found to reduce other injury crashes by 21.4% with a standard error of 3.75%.     

Multi-level Bayesian safety analysis with unprocessed Automatic Vehicle Identification data for an urban expressway

In traffic safety studies, crash frequency modeling of total crashes is the cornerstone before proceeding to more detailed safety evaluation. The relationship between crash occurrence and factors such as traffic flow and roadway geometric characteristics has been extensively explored for a better understanding of crash mechanisms. In this study, a multi-level Bayesian framework has been developed in an effort to identify the crash contributing factors on an urban expressway in the Central Florida area. Two types of traffic data from the Automatic Vehicle Identification system, which are the processed data capped at speed limit and the unprocessed data retaining the original speed were incorporated in the analysis along with road geometric information. The model framework was proposed to account for the hierarchical data structure and the heterogeneity among the traffic and roadway geometric data. Multi-level and random parameters models were constructed and compared with the Negative Binomial model under the Bayesian inference framework. Results showed that the unprocessed traffic data was superior. Both multi-level models and random parameters models outperformed the Negative Binomial model and the models with random parameters achieved the best model fitting. The contributing factors identified imply that on the urban expressway lower speed and higher speed variation could significantly increase the crash likelihood. Other geometric factors were significant including auxiliary lanes and horizontal curvature.     

Explaining two-lane highway crash rates using land use and hourly exposure

This paper describes the estimation of Poisson regression models for predicting both single and multi-vehicle highway crash rates as a function of traffic density and land use, as well as ambient light conditions and time of day. The study focuses on seventeen rural, two-lane highway segments, each one-half mile in length with varying land use patterns and where actual hourly exposure values are available in the form of observed traffic counts. Land-use effects are represented by the number of driveways of various types on each segment. Hourly exposure is represented for single-vehicle crashes as the total vehicle miles traveled and volume/capacity ratio; for multi-vehicle crashes it is the product of the hourly volumes on the main highway and the roads intersecting it along the study segment. For single-vehicle crashes, the following variables were found to be significant, with a positive or negative effect as noted: daytime (06:00–19:00 h, negative effect), the natural log of the segment volume/capacity ratio (negative), percent of the segment with no passing zones (positive), shoulder width (positive), number of intersections (negative), and driveways (mixed effects by type). Good multi-vehicle crash prediction models had quite different variables: daylight conditions from 10:00–15:00 and 15:00–19:00 h (positive), number of intersections (negative), and driveways (positive for all types). The results show that traffic intensity explains differences in crash rates even when controlling for time of day and light conditions, and that these effects are quite different for single and multi-vehicle crashes. Suggestions for future research are also given.     

Model crash frequency at highway–railroad grade crossings using negative binomial regression

Despite the fact that traffic collisions at highway–railroad grade crossings (HRGXs) are rare events, the impact of HRGX crashes is nevertheless more severe than highway crashes. Empirical studies show that traffic collisions at HRGXs are mainly attributed to railway-related and/or highway-related characteristics, particularly drivers’ abnormal behavior, driving around, or through an HRGX. These factors have different effects on crash likelihood (i.e., the number of traffic collisions or crash frequency) at an HRGX. To explore the causal relationship between crash frequency and the factors related to railroad and highway systems, we used a negative binomial regression model to identify the factors that are statistically significantly associated with traffic collisions at HRGXs, and conducted relevant sensitivity analyses to investigate the marginal effect of daily highway traffic on changes in crash frequency. The empirical study shows that the number of daily trains, the number of tracks, highway separation, annual averaged daily traffic (AADT), and crossing length had statistically significant effects on the mean number of traffic collisions (all p-values?≤?0.0487). Further, the marginal effect of the AADT on the change of crash frequency indicates that crash likelihood monotonically increases with the increase of AADT.     

The influence of economic incentives linked to road safety indicators on accidents: The case of toll concessions in Spain

The goal of this paper is to evaluate whether the incentives incorporated in toll highway concession contracts in order to encourage private operators to adopt measures to reduce accidents are actually effective at improving safety. To this end, we implemented negative binomial regression models using information about highway characteristics and accident data from toll highway concessions in Spain from 2007 to 2009. Our results show that even though road safety is highly influenced by variables that are not managed by the contractor, such as the annual average daily traffic (AADT), the percentage of heavy vehicles on the highway, number of lanes, number of intersections and average speed; the implementation of these incentives has a positive influence on the reduction of accidents and injuries. Consequently, this measure seems to be an effective way of improving safety performance in road networks.     

Models relating traffic safety with road environment and traffic flows on arterial roads in Addis Ababa

This paper presents the study carried out to develop accident predictive models based on the data collected on arterial roads in Addis Ababa. Poisson and negative binomial regression methods were used to relate the discrete accident data with the road and traffic flow explanatory variables. Significant accident predictive models were found with a number of significant explanatory variables. The results show that the existing inadequate road infrastructure and poor road traffic operations are the potential contributors of this ever-growing challenge of the road transport in Addis Ababa. The results also indicate that improvements in roadway width, pedestrian facilities, and access management are effective in reducing road traffic accidents.     

A genetic programming approach to explore the crash severity on multi-lane roads

The study aims at understanding the relationship of geometric and environmental factors with injury related crashes as well as with severe crashes through the development of classification models. The Linear Genetic Programming (LGP) method is used to achieve these objectives. LGP is based on the traditional genetic algorithm, except that it evolves computer programs. The methodology is different from traditional non-parametric methods like classification and regression trees which develop only one model, with fixed criteria, for any given dataset. The LGP on the other hand not only evolves numerous models through the concept of biological evolution, and using the evolutionary operators of crossover and mutation, but also allows the investigator to choose the best models, developed over various runs, based on classification rates. Discipulus™ software was used to evolve the models. The results included vision obstruction which was found to be a leading factor for severe crashes. Percentage of trucks, even if small, is more likely to make the crashes injury prone. The ‘lawn and curb’ median are found to be safe for angle/turning movement crashes. Dry surface conditions as well as good pavement conditions decrease the severity of crashes and so also wider shoulder and sidewalk widths. Interaction terms among variables like on-street parking with higher posted speed limit have been found to make injuries more probable.