Angle Crashes Research Articles

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).

Genetic Programming to Investigate Design Parameters Contributing to Crash Occurrence on Urban Arterials

Nonlinear models were developed to estimate crash frequency on urban arterials with partial access control. These multilane arterials consist of midblock segments joined by signalized and unsignalized intersections (or access points). Crashes included in the analysis are of three major types: rear-end, angle, and head-on. Each crash type is further sorted into mutually exclusive categories on the basis of the roadway element responsible for the crashes: midblock segment, signalized intersection, and access point. Genetic programming (GP) is adopted for predicting crash frequency. GP, which is primarily based on genetic algorithms, uses the concept of evolution to develop models through the processes of crossover and mutation. The GP modeling approach gives independence for model development without restrictions on distribution of data. The models developed were compared to the basic negative binomial models. Morning and afternoon peak periods are observed to have fewer occurrences of rear-end crashes at all roadway elements. Higher traffic volume results in an increased number of angle crashes. Instances of angle crashes have increased at signalized intersections, even at lower maximum posted speeds. A higher average truck factor increases the instances of head-on crashes on midblock segments and at signalized intersections.

Investigation of Truck-Involved Angle Crashes Based on GES and FARS Crash Databases

Truck-car angle collisions have a higher crash frequency than other truck-involved collision types and tend to increase injury severity. This paper investigates both general and fatal truck-involved angle collisions using two national crash databases (2000-2004), General estimates system (GES) and Fatality Analysis Reporting system (FARS). In this study, two-vehicle angle collisions were classified into three groups based on fault roles of truck or car drivers in the acci- dents, including Truck-Car, Car-Truck, and Car-Car crashes. The occurrence conditions of the three angle crash types were compared to each other to identify the potential risk factors such as driver characteristics, road environments, and highway designs related to the truck-involved crashes. The multinomial logistic regression is used for the statistical analy- sis. Based on the result analysis of this study, it is suggested that truck-involved angle collisions should be considered as an important scenario design for retraining or education programs for the purposes of reducing older drivers' fatality rate; improving either the conspicuity of truck trailers or lighting design of the highway would reduce the frequency and sever- ity of truck-involved angle crashes; to improve incompatibilities between truck, car, and highway design, further studies should conduct in-depth analyses of geometric factors related to driver performances and behaviors in the car-truck con- flicts at intersections.

Motorcyclist injury severity in angle crashes at T-junctions: Identifying significant factors and analysing what made motorists fail to yield to motorcycles

Evidence in literature suggested that motorists’ failure to give way to motorcycles at junctions is the main contributory factor to motorcycle–car accidents that involve gap acceptance (i.e., approach-turn and angle crashes). This paper attempts to examine how motorist’s failure to give way affects motorcyclist injury severity in angle crashes at T-junctions, while controlling for other factors (demographic, vehicle, crash, and environmental factors). Binary logistic models of motorcyclist injury severity were estimated using the data extracted from the Stats19 accident injury database (1991–2004). Angle collisions were classified into several sub-crashes based on the manoeuvres motorcycles and cars were making prior to the accidents. The modelling results showed that injuries were greatest when a travelling-straight motorcycle on the main road crashed into a right-turn car from the minor road, particularly at stop-/yield-controlled junctions. Such crash pattern was assumed to be an accident involving right-of-way violation. Using binary logistic models, factors determining the likelihood of motorist’s failure to yield to motorcycles were also examined. The implications of the research findings of this present study were provided.

Crash Analysis of Roundabouts at High-Speed Rural Intersections

Federal, state, and local agencies have identified increased intersection safety as an achievable objective in their campaigns for safer highways. Roundabouts could reduce the number of fatal and severe crashes. Roundabouts have been successful in Europe and Australia in both urban and rural environments for decades. The documented safety benefits associated with roundabouts address numerous concerns surrounding injury crashes at intersections; however, applications of roundabouts on two-lane, high-speed (40+ mph) rural highways have not been explored to the extent that urban roundabouts have in the United States. This research conducted a safety analysis of 17 intersections on high-speed rural roadways in the United States that were converted to roundabouts. The findings show the average injury crash frequency was reduced by 84%, average injury crash rate was reduced by 89%, angle crashes were reduced by 86%, and fatal crashes were reduced by 100%.

Two Low-Cost Safety Concepts for Two-Way, Stop-Controlled Intersections in Rural Areas

FHWA's Office of Safety has identified intersections as one of its safety focus areas. As part of FHWA's efforts to reduce intersection crashes and related injuries and fatalities, two low-cost concepts have been identified: (a) lane narrowing on major road intersection approaches and (b) installing channelizing islands and supplemental stop signs on minor road intersection approaches. With the cooperation of several transportation agencies, these strategies were deployed at a limited number of sites in the United States; specifically, the concepts were deployed at two-lane, two-way, stop-controlled intersections. Positive operational and safety effects were observed with the installation of the two strategies. The lane-narrowing concept was shown to significantly reduce speeds and crashes on the major road approaches. The average speed reductions for all vehicles (3.5 mph) and for trucks only (4.4 mph) were statistically significant at the 95% confidence level. Total, fatal and injury-only, and angle crashes were reduced after deployment, but only the reduction in fatal and injury-only crashes was significant at the 95% confidence level. The minor road channelizing concept is shown to improve driver compliance with traffic control and to reduce speeds and crashes on minor approaches. The average speed reductions on minor approaches (nearly 10 mph) were statistically significant at the 95% confidence level. The crash rate for total, fatal and injury-only, angle, and rear-end crashes decreased after deployment.

Incorporating Traffic Operation Measures in Safety Analysis at Signalized Intersections

Level of service (LOS) is applied to measure traffic operation conditions and is used as a performance measure in transportation planning. Safety is an important issue in operation and transportation planning. Although it is generally agreed that operational conditions have an effect on traffic safety, their relationship has not been well defined at a level of detail that adequately supports operation and planning performance measurements. Signalized intersections are the most complex locations on the roadway network and are vital to the efficiency and safety of transportation systems. Based on data collected on 164 four-legged signalized intersections in the state of Florida, this study investigated the relationship between LOS and safety at signalized intersections. The LOS was calculated for three time periods: a.m. peak, midday, and p.m. peak. The generalized estimating equations with a negative binomial link function were applied to fit crash frequency models for total crashes, rear-end and sideswipe crashes, angle and left-turn crashes, and severe crashes for the selected intersections. Among operational measures, the study found that the six-level LOS indicator was better than the average delay in the crash frequency models. Overall, the LOS indicator was a significant factor for predicting crash occurrence at intersections. However, the models showed that the operational condition itself (by only including the LOS indicator) was not sufficient to predict the safety status. LOS D was a desirable level that was associated with fewer total crashes, rear-end and sideswipe crashes, and right-angle and left-turn crashes.

Impact of “Signal Ahead” pavement marking on driver behavior at signalized intersections

When drivers are approaching a signalized intersection at the onset of yellow signal, they may hesitate to decide whether to stop or cross the intersection due to the dilemma zone. As defined in the Traffic Engineering Handbook [Institute of Transportation Engineers (ITE). (1999). Traffic engineering handbook, Washington, DC: Institute of Transportation Engineers], the dilemma zone for an intersection is a specific road segment, prior to the intersection in which a vehicle approaching the intersection during the yellow phase can neither safely clear the intersection, nor stop comfortably at the stop line. In this study, a pavement marking countermeasure is proposed to reduce the dilemma zone and improve traffic safety at signalized intersections. A pavement marking with word message ‘Signal Ahead’ that is placed on the pavement upstream of a signalized intersection may have a potential to assist drivers in stop/go decisions and improve driver performance during the signal change. This paper presents a simulator-based study that was designed to test the impact of the pavement marking countermeasure on driving behaviors, including driver’s stop/go decision, red-light running violation, brake response time, and deceleration rate of the stopping vehicle. The experiment results indicated that the marking had positive effects on drivers’ behaviors at signalized intersections. It was found that the marking can reduce the probabilities of both conservative-stop and risky-go decisions, contribute to a lower red-light running rate, and result in a lower deceleration rate for stopping drivers at higher speed limit intersections. These findings suggest that the marking countermeasure may have a potential to reduce the probabilities of both rear-end and angle crashes.

A simultaneous equations model of crash frequency by collision type for rural intersections

Safety at roadway intersections is of significant interest to transportation professionals due to the large number of intersections in transportation networks, the complexity of traffic movements at these locations that leads to large numbers of conflicts, and the wide variety of geometric and operational features that define them. A variety of collision types including head-on, sideswipe, rear-end, and angle crashes occur at intersections. While intersection crash totals may not reveal a site deficiency, over exposure of a specific crash type may reveal otherwise undetected deficiencies. Thus, there is a need to be able to model the expected frequency of crashes by collision type at intersections to enable the detection of problems and the implementation of effective design strategies and countermeasures. Statistically, it is important to consider modeling collision type frequencies simultaneously to account for the possibility of common unobserved factors affecting crash frequencies across crash types. In this paper, a simultaneous equations model of crash frequencies by collision type is developed and presented using crash data for rural intersections in Georgia. The model estimation results support the notion of the presence of significant common unobserved factors across crash types, although the impact of these factors on parameter estimates is found to be rather modest.

The significance of endogeneity problems in crash models: An examination of left-turn lanes in intersection crash models

Crash prediction models are used for a variety of purposes including forecasting the expected future performance of various transportation system segments with similar traits. The influence of intersection features on safety have been examined extensively because intersections experience a relatively large proportion of motor vehicle conflicts and crashes compared to other segments in the transportation system. The effects of left-turn lanes at intersections in particular have seen mixed results in the literature. Some researchers have found that left-turn lanes are beneficial to safety while others have reported detrimental effects on safety. This inconsistency is not surprising given that the installation of left-turn lanes is often endogenous, that is, influenced by crash counts and/or traffic volumes. Endogeneity creates problems in econometric and statistical models and is likely to account for the inconsistencies reported in the literature. This paper reports on a limited-information maximum likelihood (LIML) estimation approach to compensate for endogeneity between left-turn lane presence and angle crashes. The effects of endogeneity are mitigated using the approach, revealing the unbiased effect of left-turn lanes on crash frequency for a dataset of Georgia intersections. The research shows that without accounting for endogeneity, left-turn lanes 'appear' to contribute to crashes; however, when endogeneity is accounted for in the model, left-turn lanes reduce angle crash frequencies as expected by engineering judgment. Other endogenous variables may lurk in crash models as well, suggesting that the method may be used to correct simultaneity problems with other variables and in other transportation modeling contexts.

Assessment of freeway traffic parameters leading to lane-change related collisions

This study aims at ‘predicting’ the occurrence of lane-change related freeway crashes using the traffic surveillance data collected from a pair of dual loop detectors. The approach adopted here involves developing classification models using the historical crash data and corresponding information on real-time traffic parameters obtained from loop detectors. The historical crash and loop detector data to calibrate the neural network models (corresponding to crash and non-crash cases to set up a binary classification problem) were collected from the Interstate-4 corridor in Orlando (FL) metropolitan area. Through a careful examination of crash data, it was concluded that all sideswipe collisions and the angle crashes that occur on the inner lanes (left most and center lanes) of the freeway may be attributed to lane-changing maneuvers. These crashes are referred to as lane-change related crashes in this study. The factors explored as independent variables include the parameters formulated to capture the overall measure of lane-changing and between-lane variations of speed, volume and occupancy at the station located upstream of crash locations. Classification tree based variable selection procedure showed that average speeds upstream and downstream of crash location, difference in occupancy on adjacent lanes and standard deviation of volume and speed downstream of the crash location were found to be significantly associated with the binary variable (crash versus non-crash). The classification models based on data mining approach achieved satisfactory classification accuracy over the validation dataset. The results indicate that these models may be applied for identifying real-time traffic conditions prone to lane-change related crashes.

Safety Impacts of Photo-Red Enforcement at Suburban Signalized Intersections

Photo-red enforcement entails cameras that photograph a vehicle entering an intersection after the traffic signal has turned red; a citation is then mailed to the vehicle's owner. Such a program was operated at 13 signals in Fairfax County, Virginia, between 2000 and 2005. This paper reports how photo-red enforcement affected crash frequency at the intersections. An empirical Bayes approach was used to detect the impact of the program on crashes while main-line traffic volume, yellow time, truck percentage, number of left-turning lanes, and speed limit were controlled. Cameras were definitely correlated with decreased red light running crashes (11% to 55%), increased rear-end crashes (32% to 85%), increased total crashes (0% to 23%), and decreased red light running injury crashes (5% to 64%). Cameras were probably associated with increased total injury crashes (5% drop to a 33% raise). These results are consistent with those of other studies with one notable exception: In the case of Fairfax County, the magnitude of the increase in rear-end crashes was larger than the magnitude of the decrease in angle crashes. That difference from other studies may have been the result of the investigators’ decision to study the crash narrative and diagram instead of relying only on collision type. Because the number of induced rear-end crashes was so much larger than the number of eliminated red light running (angle) crashes, a net safety gain would be realized if and only if severity of the latter were substantially greater than severity of the former.

Examination of Fault, Unsafe Driving Acts, and Total Harm in Car-Truck Collisions

Crashes involving large trucks and passenger cars are important topics for research and countermeasure development since they represent more than 60% of all fatal truck crashes and because the passenger car occupant is much more likely to be killed. This study ( a) examined “fault” in total car–truck crashes using North Carolina Highway Safety Information System (HSIS) data for comparison with fault analyzed in previous studies of fatal crashes, ( b) used general estimates system (GES) crash data to verify unsafe driving acts (UDAs) identified by expert panels in past studies, and ( c) used North Carolina HSIS data to identify critical combinations of roadway facility type, roadway location, and crash type based on “total harm”—a measure combining both the frequency and severity of the crash. Fault in total North Carolina car–truck crashes was found to differ significantly from past fatal crash studies, with the truck driver being at fault more often than the car driver both overall and in certain crash types. Car drivers continue to be at fault much more often in head-on and angle crashes. While it was not possible to analyze all UDAs identified in prior studies, when possible, the current analyses revealed differences between the GES crash data results and the expert-based results, pointing to the need for better UDA methods if they are to be used to target treatments. Finally, using the total-harm analysis with North Carolina car–truck crashes indicated that undivided rural arterials and collectors should be primary targets for further investigation and for treatment.

Age related changes in drivers' crash risk and crash type

Age and gender differences in the rates of crash involvement of Western Australian drivers were examined using the Road Injury Database of the Road Accident Prevention Research Unit at the University of Western Australia. The population examined was all drivers of cars, station wagons and related vehicles involved in property damage, injury and fatal crashes reported to the police in Western Australia from 1 January 1989 to 31 December 1992. It was found that drivers aged <25 years were involved in ca 35% of crashes, compared with 3% for drivers 70 years and over. Drivers under 25 years had the highest rates on a population and licence basis, but when the distance travelled was taken into account, rates of crash involvement for the 75 or more age group were as high as those of the youngest age group. Females had higher rates of crash involvement than males in all age groups. Drivers over 70 years were involved in relatively more crashes involving fatalities and hospital admissions than younger drivers, although the number of such crashes was small. The youngest groups of drivers had proportionately more single vehicle crashes, drivers 30–59 years had more same direction crashes, and drivers over 60 years, particularly those over 75 years, had more direct and indirect right angle crashes. There were also age related patterns in the movements associated with these crashes. Drivers under 30 years were associated with swerving and swinging wide, drivers 30–59 years were stopped at the time of the crash disproportionately often, and drivers over 60 years were associated with turning movements. The percentage of crashes in daylight ranged from ca 64% for drivers under 20 years to a maximum of over 90% for those 80 years and over. These patterns are consistent with changes in exposure to risk of crash involvement with age, and also with changes in ability, experience and psychological function, which are also related to age.

A new decision making algorithm for airbag control

Airbag systems have been introduced to supplement the seat belt system primarily to reduce head injuries. Most of the present airbag systems use distributed mechanical sensors, which are costly, not easy to calibrate, and not effective to trigger the airbag on time for different types of crashes. An electronic sensor is more effective in the sense that the signal from the accelerometer can be digitized and analyzed to study the behavior of the signal for different types of crashes. Unfortunately, most of the algorithms which are used for electronic sensors still have some problems. They either fail to trigger the airbag on time for several types of high speed crashes such as pole and angle crashes, or if they do trigger at these crashes they also trigger at low speed barrier crashes. A new algorithm for an electronic sensor is presented. This algorithm has been developed by analyzing a large volume of data obtained from actual crash testing experiments. The algorithm works for all types of crashes. The algorithm has been tested by applying it on real crash testing data. A hardware circuit has also been proposed in the paper, which can be used to implement the algorithm for real-time operations. >

A Summary of Roadway Accidents Involving Agricultural Machinery

Four years of Ohio highway accident data (1989-1992) were analyzed for patterns in accidents involving agricultural machinery. Seventy-eight percent of two-vehicle accidents occurred during daylight hours, with a peak occurrence time interval of 3:00 P.M. to 6:00 P.M. Sideswipe and angle collisions were predominant accident types. Forty-two percent of the nighttime accidents were rear end collisions compared to 8% of the daylight crashes. Turning left was the most common action preceding sideswipe and angle crashes. Operator age, weather conditions, and alcohol use were not major factors in the majority of these accidents.