بررسی عوامل تاثیر گذار بر وقوع و شدت تصادفات از عقب Investigating Factors Affecting the Occurrence and Severity of Rear-End Crashes

۱٫ Introduction Vehicle to vehicle crashes have been classified based on the collision type into four categories: rear-end, side sweep, head-on and right angle collisions. Rear-end collisions are considered as the most frequently occurring types of crashes over the world and lead to a significant number of injuries and fatalities. For instance, in USA about one third of all crashes were rear end crashes (over 2.5 million rear end collisions are reported every year). These crashes were responsible for 30% of all injuries and fatalities (Sing, 2003). On the other side of the world, Kampen (2000) showed that about 35% of all crashes on Netherlands highways are rear-end against 9% in urban areas. In Japan, about 28% of total series crashes (i.e., crashes involving injuries or fatalities) are rear-end crashes and represent about 35% of crashes at intersections (ITARDA 1998; Wang et al., 2002). However, no information about the rear-end crashes in the middle-east region has been published before. Thus, significant efforts are needed to investigate and identify the contributing factors of rear-end crashes for more understanding of its characteristics and to develop the proper countermeasures in order to reduce its occurrence and severity. In the Emirate of Abu Dhabi AD, the Capital of UAE, the rear-end collisions represent about 17% of total severe crashes (i.e., any crash with at least one injury). Despite the decreasing trend of the traffic crashes and related fatalities between 2010 and 2014, the contribution of the rear-end crashes in the total crashes increased from 15.8% in year 2010 to 20.4 % in year 2014. In addition, the statistics showed that the severity of the rear-end crashes in AD is significantly higher (179 fatalities/1000 crashes) than the severity of other types of crashes (134 fatalities/1000 crashes). Also, it is worth mentioning that licensed drivers community in Gulf religion is significantly different than other regions in the world. For instance, in AD about 87% of the drivers are foreigners from more than one hundred different nationalities, 85% of the drivers are males and 92% are younger than 45 years old. Despite of these facts, the contributing factors that affect the occurrence and severity of the rear-end crashes have not been explicitly discussed in any prior studies. This paper mainly aims to provide extensive information and analysis about the contributing factors affect the occurrence and severity of rear-end crashes in Gulf country region based on data collected from Abu Dhabi Emirate. The investigated factors in this study include the characteristics and behavior of at-fault drivers involving in rear-end crashes, crash information, road site characteristics and weather condition. A logistic regression analysis is applied to model and investigate the impact of these factors on the rear-end crash severity. ۲٫ Literature Review Significant research effort have been undertaken to analyze the characteristics and causes of the rear-end crashes in different parts of the world. Several statistical modeling techniques were utilized to investigate the contribution factors affecting the occurrence and severity of rear-end crashes (Lao, et al. 2014). The majority of the researchers applied linear regression analysis bases in the analysis and modeling process. Abdel-Aty and Abdelwahab (2004) used the nested logit model to estimate the probability of four configurations of car-truck rear-end crashes. The results showed that driver’s visibility and inattention in the following vehicle have the largest effect on being involved in a rear-end crashes. The risk or rear-end crashes at signalized intersections was investigated by Yan, et al. (2005) using the multiple regression modeling. The model showed that seven road environment factors (number of lanes, divided/undivided highway, accident time, road surface condition, highway character, urban/rural, and speed limit), five factors related to striking role (vehicle type, driver age, alcohol/drug use, driver residence, and gender), and four factors related to struck role (vehicle type, driver age, driver residence, and gender) are significantly associated with the risk of rear-end accidents. In addition, Wang and Abdal-Aty (2006) applied negative binomial link function for risk analysis at signalized intersections. It was found that traffic volumes, speed, number of legs, right and left-turn proportion are significantly affect the occurrence of the rear-end crashes. Kim et al. (2007) estimated rear-end crash risk at freeways using a modified negative binomial regression. The results showed that urban area, curvature, offramp and merge, shoulder width, and merge section are factors found to increase rear-end crash probabilities. Harb et al. (2008) used a conditional logistic regression model to estimate rear-end crash risk in work zone. This model showed that roadway geometry, weather condition, age, gender, lighting condition, residence code, and driving under the influence of alcohol and/or drugs are significant risk factors associated with work-zone crashes. Recently, Bayesian Network BN hybrid approach has been gaining in the analysis of crash severity (Borg et al., 2014; Liang and Lee  ۲۰۱۴; Mujalli and de Oña. 2011; Zhao et al., 2012). Chen et al., (2015) developed a multinomial logit model (BN approach) to investigate the contributing factors of rear-end crash severity. The results showed that truck-involvement, inferior lighting conditions, windy weather conditions, the number of vehicles involved, etc. could significantly increase driver injury severities in rear-end crashes. The impact of the driver visibility on the vehicle collision has been investigated by Hassan and Abdal-Aty (2011). Some researchers used the nonlinear modeling approach to extract more complex interaction relationships between crash severity and the contributing factors (Wong et al., 2007; Abdel-Aty and Haleem, 2011). Lao et al. (2014), applied the generalized nonlinear modeling approach to investigate the relationship between risk of rear-end crashes and independent variable. The results showed for example that truck percentage and grade have a parabolic impact: they increase crash risks initially, but decrease them after the certain thresholds. Other approaches were applied to quantify the risk of rear-end crashes. Oh et al. (2006) utilized inductive loop detector data to determine the potential of rear-end collision based on fuzzy-clustering algorithm. The results showed that six categories were more appropriate to establish collision risk criteria. Another study in rear-end crashes used loop detectors data was conducted by Pande and Abdal-Aty (2006). It was found that the average speed and occupancy downstream has a significant contribution of rear-end crashes.. In addition, Oh and Kim (2010) utilized the trajectory data of individual vehicles to develop a risk index of the read-end collision. Meng and Qu (2012) used the time to collision TTC data collected from two road tunnel in Singapore to analysis the frequency of Rear-end crashes. Das and Abdel-Aty (2011) applied the Genetic Programing methodology to quantity the risk of rear-end crashes in urban roads. Li et al. (2014) examined the kinematic wave approach to investigate the risk index on rear-end crashes near freeway recurrent bottlenecks. It was found that the likelihood of rear-end collision is highest when the traffic approaching from upstream in near capacity state while downstream traffic is highly congested.