Risky Driving Research Articles

CNN-Based Distracted Driver Recognition: Enhancing Accuracy Through Optimized Training Techniques

Abstract. Distracted driving is a significant contributor to vehicle accidents worldwide, underscoring the need for an effective recognition model that can identify risky driver behaviors in real-time and provide timely alerts. This study proposes a Convolutional Neural Network (CNN)-based model designed to accurately detect distracted drivers. To optimize the model's performance, this paper implemented several strategic enhancements. During the preprocessing phase, 90% of the dataset was allocated for training, ensuring a comprehensive learning process, while the remaining 10% was used for validation to assess the model's accuracy. The Adam optimizer was chosen for its ability to dynamically adjust the learning rate, facilitating faster convergence to an optimal solution. Additionally, the Cross-Entropy loss function was employed to amplify errors during training, driving the model to correct inaccuracies more effectively. The model was trained over 25 epochs, resulting in an accuracy rate of nearly 80%. This level of performance demonstrates the models viability for real-world applications, where it can play a critical role in reducing accidents caused by distracted driving. Future research may focus on further refining the model by exploring advanced loss functions, optimizers, and CNN architectures, as well as incorporating more sophisticated data preprocessing techniques.

LSTM + Transformer Real-Time Crash Risk Evaluation Using Traffic Flow and Risky Driving Behavior Data

LSTM + Transformer Real-Time Crash Risk Evaluation Using Traffic Flow and Risky Driving Behavior Data

Investigating the Influence of Time Pressure on Risky Riding Behaviour among P-Hailing Riders in Malaysia: The Mediating Role of Moral Disengagement

This study investigates the impact of time pressure on risky riding behaviour among p-hailing riders in Malaysia, with moral disengagement as a mediating factor. Drawing on Bandura's Moral Disengagement Theory and the Job Demand-Resources (JD-R) Model, the research explores how time pressure influences these behaviours. A survey of 200 respondents, selected through stratified convenience sampling, was conducted across three Northern States in Malaysia, namely Perlis, Kedah, and Penang. Data analysis using Partial Least Squares Structural Equation Modelling (PLS-SEM) reveals that time pressure significantly increases risky riding behaviour, with moral disengagement significantly mediates this relationship. The study finds that younger and less experienced riders are particularly susceptible to time pressure effects. Demographic data indicate that 80% of riders are under 35 years old, with 75% having less than three years of experience. Based on these findings, the research emphasises the need for targeted interventions, including specialised training programmes and improved workload management strategies, to mitigate risks associated with p-hailing. The study contributes to a deeper understanding of the role of a cognitive mechanism factor in influencing risky riding behaviours among p-hailing riders and provides valuable insights for enhancing road safety measures. Future research directions are suggested, including exploration of additional cognitive mechanisms that may impact risky riding behaviours in this context.

Handheld Cellphone Use and Risky Driving in Adolescents

Handheld cellphone use while driving is associated with increased motor vehicle crash risk among adolescents. To examine the association of handheld cellphone use while driving with kinematic risky driving (KRD) events-hard braking and rapid acceleration-in adolescent drivers. Adolescents aged 16.50 to 17.99 years licensed 365 days or less in Pennsylvania were eligible to participate in this cross-sectional study. Enrollment occurred from July 29, 2021, to June 6, 2022. Participants downloaded a smartphone telematics cellphone app to record driving data for 60 days. Trip characteristics, including frequency, length, and duration; presence of speeding; frequency and duration of handheld cellphone use; time of day; and presence of precipitation. Kinematic risky driving events per 100 miles driven. Zero-inflated Poisson regression models examined whether individual characteristics and trip characteristics were associated with KRD. Incidence rate ratios were computed. Of 405 adolescents who responded to recruitment, 151 enrolled, 140 completed study procedures, and 119 with 12 360 trips were included in the analytic sample (60 female participants [50.4%]; mean [SD] age, 17.2 [0.4] years). Adolescents drove a mean (SD) of 103.8 (65.7) trips, 565.0 (487.3) miles, and 25.1 (19.3) hours. Adolescents had minimal night trips (1.5% [192]), and few trips with precipitation present (9.0% [1097]). Speeding occurred in 43.9% (5428) of the trips and handheld cellphone use occurred in 34.1% (4214) of the trips. Kinematic risky driving events occurred in 10.9% (1358) of the trips at a rate of 2.65 per 100 miles. In adjusted models, increased KRD events were associated with handheld cellphone use (incidence rate ratio [IRR], 2.62; 95% CI, 1.53-4.48), speeding (IRR, 2.12; 95% CI, 1.06-4.26), and minutes driving (IRR, 1.02; 95% CI, 1.01-1.02). Trips at night, precipitation presence, licensure for less than 6 months, and sex were not associated with increased KRD events. In this cross-sectional study of adolescent drivers, trips with handheld cellphone use and speeding were associated with higher rates of KRD, while individual characteristics were not. The findings suggest that smartphone telematics apps provide an opportunity to observe behaviors as well as surveil changes due to intervention efforts.

Identifying Risky Driving Behaviors through Vehicle Trajectories Collected by On-Road Millimeter-Wave Radars

Identifying Risky Driving Behaviors through Vehicle Trajectories Collected by On-Road Millimeter-Wave Radars

Interactive effects analysis of road, traffic, and weather characteristics on shared e-bike speeding risk: A data-driven approach

As electric bikes (e-bikes) rapidly develop in China, their traffic safety issues are becoming increasingly prominent. Accurately detecting risky riding behaviors and conducting mechanism analysis on the multiple risk factors are crucial in formulating and implementing precise management policies. The emergence of shared e-bikes and the advancements in interpretable machine learning present new opportunities for accurately analyzing the determinants of risky riding behaviors. The primary objective of this study is to examine and analyze the risk factors related to speeding behavior to aid urban management agencies in crafting necessary management policies. This study utilizes a large-scale dataset of shared e-bike trajectory data to establish a framework for detecting speeding behavior. Subsequently, the extreme gradient boosting (XGBoost) model is employed to identify the level of speeding risk by leveraging its excellent identification ability. Moreover, based on measuring the degree of interaction among road, traffic, and weather characteristics, the investigation of the complex interactive effects of these risk factors on high-risk speeding is conducted using bivariate partial dependence plots (PDP) by its superior parsing ability. Feature importance analysis results indicate that the top five ranked variables that significantly affect the identified results of speed risk levels are land use density, rainfall, road level, curbside parking density, and bike lane width. The interaction analysis results indicate that higher levels of road and bike lane width correspond to an increased possibility of high-risk speeding among riders. Land use density, curbside parking density, and rainfall display a nonlinear effect on high-risk speeding. Introducing road level, bike lane width, and time interval could change the patterns of nonlinear effects in land use density, curbside parking density, and rainfall. Finally, several policy recommendations are proposed to improve e-bike traffic safety by utilizing the extracted feature values associated with a higher probability of high-risk speeding.

“You are the company you keep”: A study of peer pressure on driving

Peers play an important role in shaping the behaviours of their counterparts at different stages of life. Studies have shown that peer influence/pressure has a significant role in the traffic environment as well as in many other social settings, especially for risky behaviours. Therefore, the aim of this study was to examine age and sex differences in respect of peer pressure and the role of peer pressure on driver behaviour in Türkiye. A total of 405 drivers with a mean age of 23.84 (SD=4.82) years completed an online survey consisting of the Peer Pressure on Risky Driving Scale (PPRDS), the Brief Sensation-Seeking Scale (BSSS) and the Short Driver Behaviour Questionnaire (S-DBQ, measuring aberrant and positive driver behaviour). First, the psychometric structure of the PPRDS, used for the first time in Turkish, was found to support the original structure, addressing direct pressure in the form of risk-encouraging or risk-discouraging and indirect pressure. Indirect pressure was seen to decrease with age for both males and females, whereas risk-discouraging direct pressure increased with age only for males. Sensation-seeking was positively related to aberrant behaviours. Risk-encouraging direct peer pressure and indirect pressure were positively associated with aberrant (risky) driver behaviour, whereas indirect peer pressure was associated with decreased aberrant driving behaviour and increased positive (helpful, polite) driving behaviour. The findings suggested that drivers feel strong peer pressure in traffic, but it varies by age and sex, and this pressure is stronger for young males. These findings have important implications for road safety and intervention studies.

Evaluating motorcycle Rider aggression in Mexico and Spain: Construct validity for the LatinCAM

Many researchers have delved into driver aggression for decades. Studies have been carried out in countries all over the world and attempts to curb it are well under way. However, motorcycle rider aggression has seen little effort into its description and study. The purpose of this study is to generate a scale that will allow evaluation rider aggression in Spanish-speaking countries. Through a two-phase methodology, a scale that evaluates this variable was constructed. The first phase consisted of gathering content validity for the scale, and a large group of items was comprised. Then, the second phase consisted of gathering 800 riders from Mexico and Spain, which provided the data from which construct validity was attempted and the indicators here presented were generated. Through this effort, a scale that evaluated Rider Aggression is constructed and is ready for its use in Mexico and Spain, although more research is warranted, and a larger number of Spanish-speaking countries is indicated. In the end, a definition for Rider Aggression is provided by the authors, and it was interesting that it lacks the component of Risky Driving, probably because all motorcycle use is risky in itself.

Prevalence and Profiles of Risky Driving Behavior Among US Teenagers

This survey study updates prevalence data on risky driving behaviors in a nationally representative sample of young people in the US and characterizes profiles of drivers according to the types of risky behaviors they engaged in.

Analysis and Prediction of Risky Driving Behaviors Using Fuzzy Analytical Hierarchy Process and Machine Learning Techniques

Driver behavior plays a pivotal role in ensuring road safety as it is a significant factor in preventing traffic crashes. Although extensive research has been conducted on this topic in developed countries, there is a notable gap in understanding driver behavior in developing countries, such as Pakistan. It is essential to recognize that the cultural nuances, law enforcement practices, and government investments in traffic safety in Pakistan are significantly different from those in other regions. Recognizing this disparity, this study aims to comprehensively understand risky driving behaviors in Peshawar, Pakistan. To achieve this goal, a Driver Behavior Questionnaire was designed, and responses were collected using Google Forms, resulting in 306 valid responses. The study employs a Fuzzy Analytical Hierarchy Process framework to evaluate driver behavior’s ranking criteria and weight factors. This framework assigns relative weights to different criteria and captures the uncertainty of driving thought patterns. Additionally, machine learning techniques, including support vector machine, decision tree, Naïve Bayes, Random Forest, and ensemble model, were used to predict driver behavior, enhancing the reliability and accuracy of the predictions. The results showed that the ensemble machine learning approach outperformed others with a prediction accuracy of 0.84. In addition, the findings revealed that the three most significant risky driving attributes were violations, errors, and lapses. Certain factors, such as clear road signage and driver attention, were identified as important factors in improving drivers’ risk perception. This study serves as a benchmark for policymakers, offering valuable insights to formulate effective policies for improving traffic safety.

Examining Risky Riding Behaviors: Insights from a Questionnaire Survey with Middle-Aged and Older Motorcyclists in Thailand

This research endeavors to achieve three primary objectives. Firstly, it seeks to develop a measurement model capable of assessing the motorcycle riding behavior of middle-aged individuals in Thailand. The construction of this model integrates the motorcycle rider behavior questionnaire (MRBQ) with statistical and descriptive analyses. Secondly, the research examines the accuracy of the measurement model using a factor analysis approach, comprising both exploratory and confirmatory factor analyses. Lastly, the study aims to furnish the people of Thailand with a set of guidelines for self-assessment of their motorcycle riding habits. The examination involves 399 middle-aged motorcycle riders aged 35 years or older, a significant majority of whom (81%) possess the requisite licenses for motorcycle operation, with the majority (83%) accumulating over a decade of riding experience. Through analysis, a set of 39 items capturing nuanced behaviors exhibited by middle-aged motorcyclists in Thailand is identified. These items are categorized into four distinct clusters: speed violations, control errors, traffic errors, and adherence to safety equipment protocols. The utilization of the MRBQ in this study holds significant importance, as it provides valuable insights into the riding practices of the Thai population. The resultant findings hold the potential to inform safety initiatives and strategies aimed at enhancing the overall motorcycle riding experience.

Situational Action Theory to Understand Risky Driving Behaviours in Beach Environments

This study used Situational Action Theory (SAT) to investigate personal and situational factors affecting risky driving behaviours (i.e., speeding, drink driving, drug driving) on the beach. A survey was conducted with 42 drivers, aged 17-60 years, on Teewah/Noosa Northshore Beaches. Following the SAT framework, factors including participants’ morals, attitudes, behaviours, self-control, and situational influences were analysed. Results revealed that participants generally viewed offending driving behaviour on the beach as less risky, compared to on the road. Whilst participants reported a higher likelihood of drug driving on the beach, speeding was reported as more likely on the road. Correlational analyses indicated that participants’ beliefs about driving behaviours were generally aligned between beach and road environments, potentially confirming that the ‘person’ factor within SAT, also plays a key role in beach offending. Finally, cross-tabulations suggested that the dominant motivator for beach-based risky driving was a perceived lower level of social judgment for the risky behaviours on the beach (a perception others were not concerned about risky behaviours). Preliminary findings of this study suggest that SAT could be a useful theory for understanding risky driving in distinct environments, such as beaches. SAT emphasises the interaction of personal characteristics (e.g., morals and self-control), with situational pressures (e.g., peers) and deterrents (e.g., risks). Nevertheless, factors associated with high risk driving behaviours displayed variation between beach and road contexts. Incorporating qualitative research methods might offer a more in-depth understanding of the decision-making processes underlying these risky behaviours.

Assessment of Risky Riding Behaviors Using the Motorcycle Rider Behavior Questionnaire (MRBQ) among University Students

Assessment of Risky Riding Behaviors Using the Motorcycle Rider Behavior Questionnaire (MRBQ) among University Students

A maneuver indicator and ensemble learning-based risky driver recognition approach for highway merging areas

Abstract Due to the complex traffic characteristics in highway merging areas, drivers tend to exhibit high-risk driving behaviors. To address the characteristics of driving behavior in highway merging areas, we have developed a real-time identification model for risky drivers by combining a driver risk level labeling method with load balancing-ensemble learning (LB-EL). In this paper, we explore four types of maneuver indicator indexes (MIIs)—acute direction, stomp pedal, dangerous following, and dangerous lane changing—that can describe the negative behaviors of both individual vehicles and vehicle platoons in highway merging areas. To quantize the label driver risk level, we use the interquartile range (IQR) method and Criteria Importance Though Intercriteria Correlation (CRITIC), while we evaluate the reliability of the MII using spatial analysis. Furthermore, we balance the dataset using three load balancing (LB) algorithms and create nine ensemble strategies by pairing adaptive boosting (AdaBoost), extreme gradient boosting (XGBoost), and light gradient boosting machine (LGBM) with the three LB algorithms. Finally, we validate the proposed model using trajectory data extracted from UAV videos. The results indicate that the distribution laws of risky driving behaviors in the acute direction and stomp pedal show a high degree of similarity and good matching with the distribution laws of traffic conflict points in existing research. Moreover, the SMOTE-LGBM ensemble model achieves the best performance, reaching an accuracy rate of 93.4% and a recall rate of 92.1%, which demonstrates the validity of our proposed model. This model can be widely applied to recognize risky drivers in video-based surveillance systems.

Research on Risky Driving Behavior of Young Truck Drivers: Improved Theory of Planned Behavior Based on Risk Perception Factor

Research on Risky Driving Behavior of Young Truck Drivers: Improved Theory of Planned Behavior Based on Risk Perception Factor

Real-world crash configurations and traffic violations among newly licensed young drivers with different route familiarity levels

Objective Young drivers aged 24 and below are at heightened risks of being influenced by their route familiarity levels. This study aims to compare prevalences of crash culpability, crash configurations and risky driver behaviors among newly licensed young drivers when they are driving on roads with different route familiarity levels. Methods Based on the road traffic crash and violation data in Yunnan Province of China from January 2017 through December 2019, we classified drivers’ different route familiarity levels by utilizing spatial distance away from residence-based method, including driving on high route familiarity (HRF) and low route familiarity (LRF) roads. Prevalence ratios were estimated using generalized estimating equation log-binomial regression models. Results We identified 12016 newly licensed young drivers driving on HRF roads and 2189 drivers on LRF roads. Within 48 months of licensure, young drivers on LRF roads were more likely to be at fault for their motor vehicle crashes than those on HRF roads. Young drivers on LRF roads were more likely to be with failure to obey traffic control device, with failure to yield right of way, wrong way driving, backing unsafely and improper parking compared with those on HRF roads. Drivers on LRF roads were less likely to be inattentive and driving with unsafe speed and following too closely compared with those on HRF roads. Conclusions Several basic aspects of targeted countermeasures can be put forward. Visual impacts such as rectangular rapid-flashing beacon (RRFB) can be used in order to prevent wrong way driving on the tourist roadways. Arranging safety talks and programs in colleges and universities and technical interventions like Advanced Driver Assistance Systems (ADAS) can be used to reduce young drivers’ driving distraction and overconfidence. It is recommended that the driving schools can use these research findings to include in licensure program to make young drivers more aware of the various factors that expose them to crash risks so that more defensive driving may be needed under different situations, and this can also help build the graduated driving licensure (GDL) programme in China.

Exploring the Influence of Drivers’ Visual Surroundings on Speeding Behavior

Despite awareness campaigns and legal consequences, speeding is a significant cause of road accidents and fatalities globally. To combat this issue, understanding the impact of a driver's visual surroundings is crucial in designing roadways that discourage speeding. This study investigates the influence of visual surroundings on drivers in 15 US cities using 3,407,253 driver view images from Lytx, covering 4,264 miles of roadways. By segmenting and analyzing these images along with vehicle-related variables, the study examines factors affecting speeding behavior. After filtering the images, to ensure an accurate representation of the driver’s view, 1,340,035 driver view images were used for analysis. Statistical models, including hurdle beta and bivariate probit models with random driver effects as well as Machine Learning’s eXtreme Gradient Boosting (XGBoost), were employed to estimate speeding behavior. The results indicate that factors within the driver's visual environment, weather conditions, and driver heterogeneity significantly impact speeding. Speeding behavior also varies across geographic locations, even within the same city, suggesting a connection between local context and speeding. The study highlights the importance of the driver's environment, showing that more open spaces encourage speeding, while areas with trees and buildings are associated with reduced speeding. Notably, this research differs from previous studies by utilizing real-time data from dash cameras, providing a dynamic and accurate representation of the driver's visual surroundings. This approach enhances the reliability of the findings and empowers transportation engineers and planners to make informed decisions when designing roadways and implementing interventions to address effectively excessive speeding. In addition to examining speeding behavior, the study also analyzes time-headway, a key factor affecting safety and risky driver behavior, to explore its relationship with speeding. The findings offer valuable insights into the factors influencing speeding and the driver's visual environment. These insights can inform efforts to create environments that discourage speeding (and close car following) and ultimately reduce severe accidents caused by excessive speed (and tailgating).

Towards Efficient Risky Driving Detection: A Benchmark and a Semi-Supervised Model.

Risky driving is a major factor in traffic incidents, necessitating constant monitoring and prevention through Intelligent Transportation Systems (ITS). Despite recent progress, a lack of suitable data for detecting risky driving in traffic surveillance settings remains a significant challenge. To address this issue, Bayonet-Drivers, a pioneering benchmark for risky driving detection, is proposed. The unique challenge posed by Bayonet-Drivers arises from the nature of the original data obtained from intelligent monitoring and recording systems, rather than in-vehicle cameras. Bayonet-Drivers encompasses a broad spectrum of challenging scenarios, thereby enhancing the resilience and generalizability of algorithms for detecting risky driving. Further, to address the scarcity of labeled data without compromising detection accuracy, a novel semi-supervised network architecture, named DGMB-Net, is proposed. Within DGMB-Net, an enhanced semi-supervised method founded on a teacher-student model is introduced, aiming at bypassing the time-consuming and labor-intensive tasks associated with data labeling. Additionally, DGMB-Net has engineered an Adaptive Perceptual Learning (APL) Module and a Hierarchical Feature Pyramid Network (HFPN) to amplify spatial perception capabilities and amalgamate features at varying scales and levels, thus boosting detection precision. Extensive experiments on widely utilized datasets, including the State Farm dataset and Bayonet-Drivers, demonstrated the remarkable performance of the proposed DGMB-Net.

Influence of Job Factors and Risky Riding Behaviours by Food Delivery Riders in Malaysia

Road traffic accidents in Malaysia have progressively increased, enhancing the potential of a food delivery rider breaking traffic regulations and causing a road accident. The objectives are to identify the level of safety awareness of riders when they perform their work. Next, determine the primary circumstances that contributed to the food delivery rider being involved in a road accident. Hence, identify the implications for the food delivery rider while delivering the food to the customer. It benefits food delivery riders because it minimises the number of violations and improves training for practicing road safety while delivering food.

Risky driving behavior among public transport vehicle drivers in Ethiopia using the health belief model

IntroductionEvery year, road traffic accidents cause 1.35 million deaths and up to 50 million disabilities worldwide. Low and middle-income countries account for 93% of all fatalities, despite owning 60% of the world's vehicles. Road traffic accidents contributed to a fatality rate of37 persons per 100,000 inhabitants in Ethiopia each year, with risky driving behavior accounting for 83% of traffic crashes. Using the health belief model, this study aimed to examine risky driving behavior and its associated factors among public transport vehicle drivers in East Gojjam Zone, North West Ethiopia, in 2021. MethodsA cross-sectional study was conducted among 369 public transport vehicle drivers selected by a multi-stage sampling technique. Data collected through the interviewer-administered technique were entered into Epidata version 4.6 and exported to Stata software version 14 for further statistical analysis. Multiple linear regression models were fitted with statistical significance at a p-value of 0.05 with a 95% confidence interval. ResultsA total of 369 drivers were interviewed, with a response rate of 94.1 %. The mean risky driving behavior score was 1.74 (SD + 0.36), with an expected value of a minimum of 1 and a maximum of 5, and actual minimum and maximum values of 1 and 3, respectively. Perceived severity (B = -0.23; 95 % CI (-0.37–-0.08)), perceived benefit (B = -0.58; 95 % CI (-0.68–-0.47)), perceived barrier (B = 0.23; 95 % CI (0.16–0.30)), and self-efficacy (B = -0.22; 95 % CI (-0.31–-0.13)) were statistically significant variables. The perceived benefit was the strongest predictor of risky driving behavior, and health belief model constructs explained 62.2 % of the variation in risky driving behavior. ConclusionPerceived severity, perceived benefit, perceived barrier, and self-efficacy were statistically significant variables, so corrective measures should focus on maximizing perceived severity, perceived benefit, and self-efficacy and minimizing perceived barriers.