Prediction of vehicle crashes by drivers' characteristics and past traffic violations in Korea using a zero-inflated negative binomial model

Driver's journey from historical traffic violations to future accidents: A China case based on multilayer complex network approach

Traffic violation records serve as key indicators for predicting drivers' future accidents. However, beyond statistical correlations, the underlying mechanisms linking historical traffic violations to future accidents remain inadequately understood. This study introduces a research framework to address this gap: Using Propensity Score Matching and an adapted mutual information-based feature selection algorithm to precisely identify correlations and optimal time windows between drivers' historical traffic violations and future accidents. A multilayer complex network approach was then applied to abstract and model the progression from drivers' historical traffic violations to subsequent accidents, revealing intrinsic patterns through adapted network analysis metrics and ultimately uncovering underlying mechanisms. Actual data from over 17,000 drivers in Shenzhen, China, spanning the period of 2010 to 2020, was utilized. Results revealed significant heterogeneity among driver subtypes with various driving license types regarding optimal time windows and key traffic violations indicative of future accident risks. A universal "Stable Defect Effect" was identified across all driver subtypes, characterized by persistent driving-related deficiencies resistant to temporal decay and penalties. This effect's gradual formation and maturation appear to govern the progression from traffic violations to future accidents. In addition, multilayer complex network models demonstrated significant potential in accident risk studies, particularly in providing valuable latent information by overcoming the limitations of accident data samples.

Investigating streetscape environmental characteristics associated with road traffic crashes using street view imagery and computer vision

Examining the relationship between streetscape features and road traffic crashes is vital for enhancing roadway safety. Traditional field surveys are often inefficient and lack comprehensive spatial coverage. Leveraging street view images (SVIs) and deep learning techniques provides a cost-effective alternative for extracting streetscape features. However, prior studies often rely solely on semantic segmentation, overlooking distinctions in feature shapes and contours. This study addresses these limitations by combining semantic segmentation and object detection networks to comprehensively measure streetscape features from Baidu SVIs. Semantic segmentation identifies pixel-level proportions of features such as roads, sidewalks, buildings, fences, trees, and grass, while object detection captures discrete elements like vehicles, pedestrians, and traffic lights. Zero-inflated negative binomial regression models are employed to analyze the impact of these features on three crash types: vehicle-vehicle (VCV), vehicle-pedestrian (VCP), and single-vehicle crashes (SVC). Results show that incorporating streetscape features from combined deep learning methods significantly improves crash prediction. Vehicles have a significant impact on VCV and SVC crashes, whereas pedestrians predominantly affect VCP crashes. Road surfaces, sidewalks, and plants are associated with increased crash risks, while buildings and trees correlate with reduced vehicle crash frequencies. This study highlights the advantages of integrating semantic segmentation and object detection for streetscape analysis and underscores the critical role of environmental characteristics in road traffic crashes. The findings provide actionable insights for urban planning and traffic safety strategies.

Identification and screening of key traffic violations: based on the perspective of expressing driver's accident risk

Drawing on the core idea of Propensity Score Matching, this study proposes a new concept named Historical Traffic Violation Propensity to describe the driver's historical traffic violations, and combines the new concept with an improved mutual information-based feature selection algorithm to construct a method for screening key traffic violations from the perspective of expressing driver's accident risk. The validation analysis based on the real data collected in Shenzhen demonstrated that drivers' state of Historical Traffic Violation Propensity on 19 key traffic violations screened have a stronger predictive ability of their subsequent accidents compared to the level in existing research. The positive state of Historical Traffic Violation Propensity on 'Drinking', 'Parking in dangerous areas', 'Wrong use of turn lights', 'Violating prohibited and restricted traffic regulations', and 'Disobeying prohibition sign' will increase the probability of a driver's subsequent accident by more than 1.7 times. The research provides directions to more efficiently and accurately capture the driver's accident risk through historical traffic violations, which is valuable for identifying high-risk drivers as well as the key psychological or physical risk factors that manifest in daily driving activities and lead to subsequent accidents.

Predicting Risky and Aggressive Driving Behavior among Taxi Drivers: Do Spatio-Temporal Attributes Matter?

Risky and aggressive driving maneuvers are considered a significant indicator for traffic accident occurrence as well as they aggravate their severity. Traffic violations caused by such uncivilized driving behavior is a global issue. Studies in existing literature have used statistical analysis methods to explore key contributing factors toward aggressive driving and traffic violations. However, such methods are unable to capture latent correlations among predictor variables, and they also suffer from low prediction accuracies. This study aimed to comprehensively investigate different traffic violations using spatial analysis and machine learning methods in the city of Luzhou, China. Violations committed by taxi drivers are the focus of the current study since they constitute a significant proportion of total violations reported in the city. Georeferenced violation data for the year 2016 was obtained from the traffic police department. Detailed descriptive analysis is presented to summarize key statistics about various violation types. Results revealed that over-speeding was the most prevalent violation type observed in the study area. Frequency-based nearest neighborhood cluster methods in Arc map Geographic Information System (GIS) were used to develop hotspot maps for different violation types that are vital for prioritizing and conducting treatment alternatives efficiently. Finally, different machine learning (ML) methods, including decision tree, AdaBoost with a base estimator decision tree, and stack model, were employed to predict and classify each violation type. The proposed methods were compared based on different evaluation metrics like accuracy, F-1 measure, specificity, and log loss. Prediction results demonstrated the adequacy and robustness of proposed machine learning (ML) methods. However, a detailed comparative analysis showed that the stack model outperformed other models in terms of proposed evaluation metrics.

Effects of driver sleepiness and fatigue on violations among truck drivers in India

This study aims at capturing the influence of driver drowsiness on prevalence of traffic violations among long-haul truck drivers. The study is based on a roadside survey of 453 long-haul truck drivers, stopping at eateries and rest locations on highways connected to three Indian cities- Mumbai, Indore and Nagpur. The survey questionnaire was categorized into three sections: driver demographics, work-rest schedules and safety critical driver behavior (violations and lapses) in the last five years. The questions regarding unsafe driving practices like speeding, overtaking were combined to form a single factor 'violations' using Principal Component Analysis (PCA). A generalized linear model using negative binomial regression predicted young drivers (aged below 25 years), long working hours, insufficient sleeping hours, driving after mid-night, sleepiness on the wheel and frequent traffic violations as significant contributors of violations among the long-haul truck drivers.

Experience as a Safety Factor in Driving; Methodological Considerations in a Sample of Bus Drivers

Experience is generally seen as an important factor for safe driving, but the exact size and details of this effect has never been meta-analytically described, despite a fair number of published results. However, the available data is heterogeneous concerning the methods used, which could lead to very different results. Such method effects can be difficult to identify in meta-analysis, and a within-study comparison might yield more reliable results. To test for the difference in effects between some different analytical methods, analyses of data on bus driver experience and crash involvement from a British company were conducted. Effects of within- and between-subjects analysis, non-linearity of effects, and direct and induced exposure methods were compared. Furthermore, changes in the environmental risk were investigated. Between-subject designs yielded smaller effects as compared to within-subjects designs, while non-linearity was not found. The type of exposure control applied had a strong influence on effects, as did differences in overall environmental risk between years. Apparently, “the effect of driving experience” means different things depending upon how calculations have been undertaken, at least for bus drivers. A full meta-analysis, taking several effects of methodology into account, is needed before it can be said that the effect of driving experience on crash involvement is well understood.

Effects of socio-demographic, personality and mental health factors on traffic violations in Chinese bus drivers

The present study aims to determine the association between bus drivers' socio-demographic characteristics, personality traits, mental health and traffic violations. This case-control study included 596 bus drivers who were recruited during October 2014 to May 2016, including 295 drivers with traffic violations and 301 drivers without traffic violations. The bus drivers' personality traits and mental health were assessed by the Eysenck Personality Questionnaire (EPQ) and the Symptom Checklist (SCL-90-R). Drivers aged 26-35 years were 72% less likely to be involved in traffic violations compared to drivers aged ≤25 years (OR:0.284,95%CI:0.137-0.586). Drivers with ≤2 years driving experience were associated with almost a three-fold increased risk of traffic violations compared to ≥21 years driving experience (OR:3.174,95%CI:1.097-9.187). The OR value decreased with the increase of annual income (OR:4.631,95%CI:2.667-8.042;OR:3.569,95%CI:2.038-6.251;OR:3.781,95%CI:1.999-7.151). Occasionally drinking drivers and regularly drinking drivers, compared to nondrinking drivers, exhibited a higher risk of traffic violations (OR:2.487,95%CI:1.521-4.065;OR:3.271,95%CI:1.387-7.716).Extroversion and neuroticism were identified as significant factors associated with traffic violations (OR:1.262,95%CI:1.145-1.393;OR:1.159,95%CI:1.060-1.267).Somatization increased eleven-fold risk of bus drivers' traffic violations (OR:11.185,95%CI:4.563-27.419). The results revealed that bus drivers' traffic violations were mainly affected by specific socio-demographic characteristics, personality traits and mental health, which increase the risk of traffic violations.

Predicting Future Driving Risk of Crash-Involved Drivers Based on a Systematic Machine Learning Framework

The objective of this paper is to predict the future driving risk of crash-involved drivers in Kunshan, China. A systematic machine learning framework is proposed to deal with three critical technical issues: 1. defining driving risk; 2. developing risky driving factors; 3. developing a reliable and explicable machine learning model. High-risk (HR) and low-risk (LR) drivers were defined by five different scenarios. A number of features were extracted from seven-year crash/violation records. Drivers’ two-year prior crash/violation information was used to predict their driving risk in the subsequent two years. Using a one-year rolling time window, prediction models were developed for four consecutive time periods: 2013–2014, 2014–2015, 2015–2016, and 2016–2017. Four tree-based ensemble learning techniques were attempted, including random forest (RF), Adaboost with decision tree, gradient boosting decision tree (GBDT), and extreme gradient boosting decision tree (XGboost). A temporal transferability test and a follow-up study were applied to validate the trained models. The best scenario defining driving risk was multi-dimensional, encompassing crash recurrence, severity, and fault commitment. GBDT appeared to be the best model choice across all time periods, with an acceptable average precision (AP) of 0.68 on the most recent datasets (i.e., 2016–2017). Seven of nine top features were related to risky driving behaviors, which presented non-linear relationships with driving risk. Model transferability held within relatively short time intervals (1–2 years). Appropriate risk definition, complicated violation/crash features, and advanced machine learning techniques need to be considered for risk prediction task. The proposed machine learning approach is promising, so that safety interventions can be launched more effectively.

Do Silver Zones reduce auto-related elderly pedestrian collisions? Based on a case in Seoul, South Korea

Inaugurated in 2007, in Seoul, South Korea, the Silver Zone is a designated pedestrian safety zone for the elderly that adopts speed limit measures such as traffic signage and road surface markings. In this study, we empirically investigate the effectiveness of the Silver Zone in two respects: first, whether the establishment of the Silver Zone has lowered the number of elderly pedestrian collisions, and second, whether Silver Zones are established in the appropriate areas, that is, those with the highest frequency of such collisions. From our quasi-experimental statistical analysis, Difference-in-Difference, we learn that the Silver Zone has no effects on reducing elderly pedestrian collisions. From our spatial statistical analyses-Kernel Density mapping and Bivariate Moran's I-we found a spatial mismatch between the frequency of senior pedestrian-vehicular collisions and the location of Silver Zones. For better performance of the Silver Zone system, we suggest additional types of physical measures to be integrated into the Silver Zone system. Municipal-level comprehensive master plan for Silver Zone system is also necessary, under which local governments should use periodic surveys to inventory and prioritise the locations of highest elderly pedestrian-vehicular collisions.