Accident Factors Research Articles

Analyzing Crash Severity: Human Injury Severity Prediction Method Based on Transformer Model

Traffic accident-related injuries and fatalities are a serious global public health and social development challenge. The accurate prediction of crash severity improves road safety and reduces casualties, as well as serving road managers and policy makers. Prediction models need to learn and analyze the various characteristic factors of traffic accidents and capture from them the inherent complex relationship between accident characteristics and the severity of traffic accidents. However, most accident prediction studies lack analytical predictions of injury severity, and predictive models rely on the content and quality of accident datasets. To increase the robustness and accuracy of prediction models, this paper leverages a Transformer-based architecture for the severity prediction of traffic collisions from human injury severity. This framework learns both text and sequence data from accident datasets. After comparative analysis, the framework can achieve the prediction of human injury severity under different data categories and show good prediction performance at low injury severity levels using only textual data or sequence data.

A New Framework for Real-Time Detection of Loss of Tail Rotor Effectiveness

Loss of tail rotor effectiveness (LTE), also known as unanticipated yaw, has been recognized as a major contributing factor in several helicopter accidents where directional control is lost. Through the helicopter flight data monitoring (HFDM) program, pilots receive constant flight evaluation reports to support LTE risk mitigation. Nevertheless, the existing LTE safety metric presents several pitfalls that hinder the reliability of the detection of violations of boundaries for flight safety. This paper presents a new framework for the development and evaluation of an expanded safety metric to enhance the detection of proximity to LTE within the HFDM program. The new metric combines three different underlying phenomena that can lead to an LTE event, that is, loss of weathercock stability, running out of pedal (tail rotor collective) for trim, and tail rotor vortex ring state. A parametric surrogate model for the detection of proximity to an LTE event is formulated by combining results from physics-based simulations with machine learning techniques. The new framework allows for operator fine-tuning of proximity to LTE boundaries, and, hence, it provides flexibility in risk management from postflight analyses. Finally, the framework is compared against the method currently used within the HFDM program, confirming its enhanced detection of the proximity to LTE.

Characteristics exploration of human and organizational factors in maritime traffic accidents

Characteristics exploration of human and organizational factors in maritime traffic accidents

Analysis and machine-learning-based prediction of beach accidents on a recreational beach in China

AbstractBeachgoers are sometimes exposed to coastal hazards, yet comprehensive analyses of characteristics and potential factors for beach accidents are rarely reported in China. In this study, information on beach accidents was collected on a recreational beach from 2004 to 2022 by searching the web or apps. The characteristics of beach accidents were therefore analysed in terms of age, gender, and activity of beachgoers. The potential factors were resolved in environmental aspects of meteorology, waves, tides, and beach morphology. Results show that beach accidents mainly occur in summer, with the highest occurrence in the afternoon and evening. The number of male beachgoers in accidents is five times higher than that of females. 90% of accidents occur when the beach is at a high-risk level for rip currents, providing evidence for the accuracy of the risk map built in a previous study. Three machine learning models, i.e., Support Vector Machine, Random Forest, and BP Neural Networks, are trained to predict beach accidents. The performances of these three machine learning algorithms are evaluated in terms of precision, recall, and F1 score. Support Vector Machine and BP Neural Networks significantly outperform Random Forest in terms of prediction. The accuracy in predicting "safe" and "dangerous" classes is approximately 80% of the Support Vector Machine model. This paper provides a preliminary study of machine learning based beach accident prediction for a specific tourist beach. In the future, machine learning will be applied to predict tourist beach accidents throughout mainland China.

Modeling the traffic accident probability of Urban roads based on an improved Bayesian network

ABSTRACT Urban road traffic accidents are severely threatening the safety of human life and property. In this study, the Random Forest (RF) algorithm was used to identify the significant risk factors of road section accidents and intersection accidents, and a probability prediction model of urban road traffic accidents based on the improved Bayesian network (IBN) was constructed. Next, a method is proposed to identify the accident-prone points on urban roads. The study results showed that: (1) There are significant differences in the factors influencing the probability of accidents at road sections and intersections. (2) When different combinations of influencing factors change, the probability of accidents at road sections and intersections also changes. Finally, based on the data from road sections and intersections in Tianjin, accident probability thresholds of 10.28% and 6.69% respectively have been determined, which can accurately identify the accident-prone points on urban roads.

Factors contributing to road traffic accidents in the Gaza Strip a comprehensive analysis

Road traffic crashes (RTCs) are considered one of the major public health issues in many countries worldwide. Investigating factors of traffic crashes, accidents, and disasters can facilitate and aid in identifying measures to mitigate their frequency and severity as well as occurrence and impact, thereby enhancing road safety. This study aims to investigate the factors that contribute to road traffic accidents in the Gaza Strip, Palestine. A comprehensive questionnaire, comprising four sub-groups of factors: human, vehicle, road, and environmental, collected data from a sample of respondents, including drivers, passengers, pedestrians, and other road users. The study utilized the relative importance index (RII) to analyze and classify each factor group after analyzing 385 questionnaires in SPSS. Over 83% of respondents across all age categories identified overspending as the most significant factor causing road traffic crashes, according to the study’s findings. Additionally, more than 78% of respondents believed that weak braking or brake failure was the highest and most significant vehicle-related factor contributing to road traffic crashes. Among the road factors, 75.4% of respondents identified low or no street illumination as a significant contributor to road traffic crashes, ranking it as the fourth most influential element overall, while environmental factors had the least impact compared to other variables.

Enhanced identification of equipment failures from descriptive accident reports using language generative model

PurposeEquipment failure is a critical factor in construction accidents, often leading to severe consequences. Therefore, this study addresses two significant gaps in construction safety research: (1) effectively using historical data to investigate equipment failure and (2) understanding the classification of equipment failure according to Occupational Safety and Health Administration (OSHA) standards.Design/methodology/approachOur research utilized a multi-stage methodology. We curated data from the OSHA database, distinguishing accidents involving equipment failures. Then we developed a framework using generative artificial intelligence (AI) and large language models (LLMs) to minimize manual processing. This framework employed a two-step prompting strategy: (1) classifying narratives that describe equipment failures and (2) analyzing these cases to extract specific failure details (e.g. names, types, categories). To ensure accuracy, we conducted a manual analysis of a subset of reports to establish ground truth and tested two different LLMs within our approach, comparing their performance against this ground truth.FindingsThe tested LLMs demonstrated 95% accuracy in determining if narratives describe equipment failures and 73% accuracy in extracting equipment names, enabling automated categorical identifications. These findings highlight LLMs’ promising identification accuracy compared to manual methods.Research limitations/implicationsThe research’s focus on equipment data not only validates the research framework but also highlights its potential for broader application across various accident categories beyond construction, extending into any domain with accessible accident narratives. Given that such data are essential for regulatory bodies like OSHA, the framework’s adoption could significantly enhance safety analysis and reporting, contributing to more robust safety protocols industry-wide.Practical implicationsUsing the developed approach, the research enables us to use accident narratives, a reliable source of accident data, in accident analysis. It provides deeper insights than traditional data types, enabling a more detailed understanding of accidents at an unprecedented level. This enhanced understanding can significantly inform and improve worker safety training, education and safety policies, with the potential for broader applications across various safety-critical domains.Originality/valueThis research presents a novel approach to analyzing construction accident reports using AI and LLMs, significantly reducing manual processing time while maintaining high accuracy. By identifying equipment failures more efficiently, our work lays the groundwork for developing targeted safety protocols, contributing to overall safety improvements in construction practices and advancing data-driven analysis processes.

Modelling the risk factors for accidents in recreational divers: results from a cross-sectional evaluation in Belgium.

Characterisation of the recreational diving community could help to identify scuba divers at risk for accidents. We performed a cross-sectional evaluation in a federation for recreational scuba divers in Belgium (Duiken.Vlaanderen). Using binary logistic regression, factors predictive for accidents leading to hospitalisation were identified. Of the 710 members, 210 (29.6%) participated in the survey, representing 140,133 dives. Age was > 50 years in 55% and the median (interquartile range [IQR]) number of dives was 380 (IQR 140-935). Cardiac (9.5%), orthopaedic (11.0%), ear-nose-throat (ENT, 10.5%) and allergic diseases (30.5%) were the top four morbidities. Twenty percent reported taking cardiovascular medication. Decompression accidents, barotrauma of the ear and musculoskeletal injuries were reported in 11.0, 11.9 and 11.0%. Fifty-five divers (26.2%) reported incidents not necessitating a medical intervention. For 36 divers (17.1%), medical interventions were necessary. Among these, 13 divers (6.2%) were hospitalised at least once and 12 (5.7%) of these needed hyperbaric oxygen therapy (HBOT). The absolute risk for hospitalisation or HBOT was 0.01% per dive. Age, advanced diving qualification, more dives annually, cardiac or ENT pathology and cardiac medication were significantly associated with an increased likelihood of hospitalisation resulting from diving accidents. In a multivariate risk model, ENT comorbidity (odds ratio [OR] 9.3; P = 0.006) and cardiac medication (OR 5.6; P = 0.05) predicted hospitalisation due to a diving accident. One in six recreational scuba divers required a medical intervention at least once during their career, while 6.2% were hospitalised or received HBOT. Ear nose and throat comorbidity and cardiac medication were strong predictors for accidents. These should be given sufficient weight in dive medical examination.

Evaluation of the impact of landscape zone on driver distraction and visual comfort in road tunnels

Studies have shown that accidents in tunnels can be more severe than those in open road networks, and drivers are considered the critical factor for such accidents. An in-depth understanding of driver behavior helps analyze pre-accident safety measures and develop policies. Decoration of landscape zones is considered an effective measure for improving the driving environment inside road tunnels and has been widely adopted in China. However, the research on the impact of landscape zones on driving safety and comfort remains inadequate. It is worthwhile to explore whether the rich and colorful decoration of landscape zones results in driver distraction by excessively capturing driver’s attention, thereby increasing accident risk. To investigate this issue, a driving simulator was used, and tunnel interior scenarios characterized by three landscape zones with different spatial distributions (namely “wall,” “vault,” and “round”) were established. Eye tracking, driving performance, and pupil diameter data were collected to further analyze the driver’s psychological and physiological conditions while driving. The results showed that i) compared to no-landscape zones, the number of driver distractions increased when driving through landscape zones, with distractions caused by vault and round landscapes being particularly noticeable; ii) wall landscapes have certain advantage in maintain longitudinal driving speeds of vehicles more stable; iii) using normalized pupil diameter as an indicator for evaluating visual sensation, all three landscape zones can effectively improve the driver’s visual comfort, with no significant differences among the three; and iv) the pupil diameter changes of the wall landscapes were relatively stable compared to others, indicating the preference for landscape decoration on sidewalls inside the tunnels. The results of this investigation provide new insights into measures for improving the driving environment inside the road tunnels.

FACTORS RELATED TO WORK FATIGUE IN OCBC TOWER CONSTRUCTION PROJECT FINISHING WORKERS AT PT BUANAMAJU HASTASENTOSA IN 2024

Work fatigue is a problem in the field of Occupational Health and safety that can be one of the factors in workplace accidents. In 2018, there were 173,105 cases in the management and construction industry. With one of the factors causing the occurrence is the condition of fatigue at work. In 2013, and 65% of the workforce experienced physical fatigue. This study aims to determine the factors associated with work fatigue at PT Buanamaju Hastasentosa workers where finishing work is done at night because it is done when employees go home from work and many construction workers are overtime because the work has not been completed. The population of this study is all workers of PT Buanamaju Hastasentosa as many as 60 workers who work in the finishing project OCBC tower South Jakarta. Data analysis of this study using the chi square assay α =0.05. The results showed that workers who experience severe fatigue there are 24 (40%) workers there and the calculation of the bivairiat test obtained age (p-value=0.023), smoking status (p-vaue=0.035), sleep quality (p-value=0.027), between periods of work (p-value=0.087). Conclusion in this study there is a relationship between age, smoking status and sleep quality with work fatigue and there is no relationship between the length of work with work fatigue. The author's suggestion is that there are still workers who experience severe fatigue, to reduce fatigue, the company can provide training to prevent work fatigue.

Investigation on Mechanical Shock Wave Protective and Thermodynamic Properties of SiO2-Aerogel-Modified Polyurea.

In recent years, industrial explosion accidents are frequent, causing serious negative influences on society. Mechanical shock waves, as a typical destructive factor in explosion accidents, can cause serious personal injury and building damage. In addition, actual explosion accidents usually involve heat sources, harming protective materials and personnel. In this study, we designed SiO2-aerogel-modified polyurea and studied the effects of manufacturing pressure process and the concentration of SiO2 aerogel on the mechanical shock wave mitigation and thermodynamic properties of the modified polyurea. The results show that the addition of SiO2 aerogel can improve the mechanical shock wave mitigation performance of polyurea. The maximum peak overpressure and acceleration mitigation rate of the material has reached 17.84% and 62.21%, respectively. The addition of SiO2 aerogel helps to reduce the thermal conductivity of materials and improve the thermal insulation performance, and the atmospheric pressure process is more conducive to improving the thermal insulation performance of materials. The minimum thermal conductivity of the material has reached 0.14174 W/m·K, which is 45.65% lower than that of pure polyurea. The addition of SiO2 aerogel has different effects on the limiting oxygen index (LOI) of polyurea. Using a vacuum process, the LOI value increased with the increase in the SiO2 aerogel concentration, while using atmospheric pressure, the LOI value increased but is always lower than 21% and lower than pure polyurea. Thermogravimetric analysis showed that the addition of SiO2 aerogel under the vacuum process was helpful to improve the thermal stability of materials. However, atmospheric pressure would disrupt the thermal stability, manifested in a decrease in peak degradation temperature, an increase in peak degradation rate, and a decrease in residual mass.

Research on Natural Gas Leakage and Explosion Mechanisms in a Container House

As unconventional building structures, container houses are now widely used in urban tourism to create characteristic buildings. Nowadays, natural gas accidents occur frequently in cities and towns; however, the development of laws and influencing factors of natural gas accidents in container buildings have rarely been studied. In this paper, a natural gas explosion test was carried out in an ordinary container house, and a numerical simulation was carried out according to the test results. The influence of methane proportion, ignition position, pressure-relief area, and pressure-relief intensity on the explosion load was analyzed. Research shows that natural gas will gather from top to bottom during the process of leakage and diffusion, and vertical stratification will occur. The most unfavorable working condition is 9.5% methane. Using the roof of the container house as a pressure-relief panel can effectively control the influence range of natural gas explosion accidents and help reduce accident losses. It is suggested that the stacking of container buildings should be reduced as much as possible, and the roof strength should be weakened to ensure structural safety. The research results have certain reference values for the disaster prevention and reduction design of urban characteristic buildings.

Machine Learning-based Prediction Analysis of Potential Factors in Traffic Accidents

Every year, traffic accidents cause a great deal of death, serious injury, and financial damage, making it a major global problem. Predicting accident severity is crucial for implementing effective traffic management and safety strategies. This study employs Random Forest and Deep Neural Network (DNN) models and so on from machine learning algorithms to predict the traffic accidents severity, identifying related outcomes from a joint dataset collected by the Seattle Department of Transportation as well as UKs Department for Transport. The results showcase that snowy road surfaces, standing water and dusk or insufficient lighting conditions causes respectively an excess of 10.75%, 10.44% and 13.01% above the average number of causalities in traffic accidents. The DNN model achieved the highest accuracy (91.12%), outperforming other methods by 310 percentage, especially in severe crash detection performance. In contrast, Random Forest model offers better stability across multiple classification threshold, illustrating better performance on general tasks. By diagnosing the high-risk conditions, traffic authorities can implement tailored interventions, ranging from boosting road maintenance efforts during adverse weather, improving street lighting, and increasing safety measures during peak times, to boost the standard of protection. Future research should investigate the method to optimise the robustness and generalization of DNN model through techniques like threshold adjustment, or adopting ensemble learning mechanisms as well as devising new segments that are more relevant.

Tensor-based analysis of eclectic-reported crowd accidents: An innovative retrospective methodology design study

IntroductionAnalysis of crowd accidents contributes to accident prevention. Therefore, we employ a tensor-based approach. The innovative tensor-based approach facilitates the streamlining of longitudinal studies, promotes error detection, and enhances the transparency and traceability of data collection. This study focuses on crowd accidents, the direct cause of which is the movement of the crowd (Excluding other external factors: e.g. fire, structural damage.). It aims at investigating the reliability of the records documented in relation to crowd accidents and the type of repetitions that can be found in the events. Materials and methodsThe study employed a web-based retrospective methodology with innovative tensor-based analysis, examining 186 fatal crowd accidents from 1979 to 2023. Data was collected from public sources, including news reports, government reports, and scientific publications. The analysis considered the following variables: event type, place, date, number of victims, cause, environmental characteristics, date and reliability of documented information source origination. Tensor-based method combines the improvement of the quality of the coverage and investigates changes in content over time. The seven-step method, which stores information about crowd accidents in matrices, is presented here in detail. The vcr factor is introduced to evaluate the credibility of sources. ResultsThe results show that those news items about crowd accidents are the most reliable which were created 2 years after the events. Crowd accidents are analyzed based on their influencing defining characteristics. We claim that we were able to isolate new risk factors related to the locations of crowd accidents. Globally, we focus on accidents that occurred during donation distributions and when entering buildings. ConclusionIt can be concluded that the new, seven-step, tensor-based data collection method improves the credibility value of individual information by more than 25 %. The impact of accident factors plays a key role in establishing risk factors and in the prevention of accidents. The tensor-based approach can be directly applied to record databases, enhance data provenance, and capture the temporal evolution of information.

An emergency linkage system of urban gas pipeline network based on Bayesian network

The occurrence of urban gas pipeline network accidents will pose a great threat to people's lives and property. To protect the lives and property of urban residents and ensure the normal operation of the city, this paper studied the influencing factors of urban gas pipeline network accidents. Firstly, the influencing factors of urban gas pipeline network accidents were summarized and the corresponding fishbone diagram was constructed following the data from the U.S. Department of Transportation Pipeline and Hazardous Materials Safety Administration. Secondly, based on the determined accident number model, the corresponding Bayesian network model was constructed to calculate the state probability and the sensitivity value. Finally, according to the relevant information of urban gas pipeline network management, combined with the numerical results of priori probability and sensitivity, an emergency linkage system based on the urban gas pipeline network was constructed. The calculation results of state probability and the sensitivity analysis indicate that corrosion is the most influential factor contributing to accidents in the urban gas pipeline network. The temperature change has the greatest influence on the occurrence of urban gas pipeline network accidents. Special attention should be paid to the prevention and management of these two aspects of the emergency linkage system. Besides, in order to verify the effectiveness of the system, this paper also uses AHP to evaluate the application of the system. This study aims to provide valuable insights for enhancing the safety management of urban gas pipeline networks, thereby elevating their safety management standards.

Latent class analysis of autonomous vehicle crashes

Introduction: Since September 2014, the California Department of Motor Vehicles has requested autonomous vehicle (AV) manufacturers to report their accidents if they take field tests on public roadways in California. These collision reports are heterogeneous containing a variety of accident factors. Method: To describe the accident more elaborately, we add three new category variables: ‘traffic control and status,’ ‘speed/speed change,’ and ‘type of accident location,’ extracted from crash narratives. Combining with the existing variables as model inputs, we use Latent Class Analysis (LCA) to investigate the mixture types of traffic accidents. After using ‘Mplus’ (LCA tool), the data set with 308 cases has been segmented into three clusters, including ‘rear-end collisions after the speed change of AV,’ ‘sideswipe collisions at parking places,’ and ‘hit-object collisions in normal traffic road.’ Results: These three clusters are not highlighted in previous literature and Cluster 1 shows AV should not be designed too ethically. To follow the driving habits of traditional drivers, AVs should accelerate vehicles quickly when they start to move and delay stopping in front of stop lines, traffic lights, and yielding. The cluster-based analyses show that applying LCA as a preliminary analysis can reveal the interesting hierarchical patterns hidden in the dataset and help traffic safety researchers improve AV safety performances.

A data-driven ISM-BN model for safety analysis of inland shipping in the Pearl River Basin

Inland shipping of the Pearl River plays an important role in the Chinese shipping system. To ensure navigation safety, we collect reports of maritime accidents from 2015 to 2022 in the Pearl River basin. This article extracts influencing factors by collecting the experience of inland waterway safety navigation and analyzing accident reports. Then, this paper uses the interpretative structural modeling method (ISM) to build a correlation model. Using a data-driven Bayesian network (BN), it analyzes the impact of various factors on the safety navigation in the Pearl River. The model validation is completed by compared with tree augmented naive Bayes classifiers (TAN) network using the same validation samples, through validation with the test set, the prediction accuracy has improved by 25%. The results indicate factors such as vessel type, accident month, accident day and time, etc. have a significant impact on the safety of navigation in the inland Pearl River waterway. The method used can identify important risk factors for accidents and the average predictive probability of validation samples reaches 87.03%. These research results could be extended to maritime management efforts in the Pearl River Basin.

Investigation of the risk influential factors of maritime accidents: A novel topology and robustness analytical framework

This study aims to develop a novel and fully data-driven approach to analyse the maritime accidents risk influential factors (RIFs) by integrating Association Rule Mining (ARM) and Complex Network (CN) modelling. Firstly, a comprehensive dataset comprising 21,206 maritime accident records from Marine Accident Investigation Branch and Transportation Safety Board is collected and processed to serve as the foundational data source supporting the development of the new approach. Secondly, a novel Combined Association Rule Mining method is proposed to extract the interconnections among RIFs, with the mined results mapped into a CN framework. Finally, two importance ranking algorithms, namely the PageRank-Information-Entropy algorithm and edge betweenness centrality, are applied to identify the key RIFs and their information transmission paths. By simulating deliberate and random attacks within networks, a robustness analysis is conducted to further explore the evolution of RIFs. The findings reveal that ship-related factors demonstrate greater centrality and connectivity, exerting a more substantial influence on information propagation within the network structure. The robustness analysis illustrates that strategic node and edge removals are effective in preventing risk propagation. It therefore makes contributions to the development of a theoretical basis for stakeholders to develop cost-effective preventive measures against specific RIFs, ultimately enhancing maritime safety.

Gerenciamento de Recursos da Tripulação - Fator de Prevenção de Acidentes

The article "Crew Management: Accident Prevention Factor" discusses the evolution of Crew Resource Management (CRM) in aviation, emphasizing its relevance in reducing air accidents. Since the 1980s, CRM has been implemented to enhance communication and collaboration among crew members, aiming to minimize human error, which is a major factor in accidents. The text discusses the different generations of CRM, from the first, focused on the cockpit, to the fourth, which encompasses the entire crew and external resources. Despite the initiatives of the Air Force Command, such as courses and training, there are still challenges in standardizing and adopting CRM across all air units. The article concludes that including CRM in the training curriculum for aviators is essential to disseminate this doctrine and reduce the occurrence of human errors.

Accident Prevention System for Cars, Buses, Trucks

Driver drowsiness is a significant factor in road accidents, contributing to a large number of crashes annually. To enhance road safety, the development of a Drivers Drowsiness Detection System is essential. This system builds to detect early signs of lazyness in drivers and provide timely alerts to prevent accidents. The proposed system uses to track key indicators of drowsiness, such as eye closure, head position, yawning, and blink rate. By utilizing a camera-based real-time monitoring system, it continuously analyzes the driver’s facial expressions. The system employs cadcade algorithms to distinguish a featurebased object detection algorithm to detect objects from images between alert and drowsy states based on collected data. When drowsiness is detected, the system triggers alerts like audio warnings, seat vibrations, or even vehicle slowing mechanisms, depending on the integration. This technology has the potential to significantly reduce the number of road accidents related to driver fatigue and could be applied in both personal vehicles and commercial fleets. The implementation of such systems represents a critical step toward improving road safety and decreasing driver-related incidents. Nowadays it is very challenging to stay active all the time due to busy schedules. Falling asleep while driving can lead to serious consequences, accidents, and even death. This situation is much more common and therefore it is very important to solve this problem. So, to solve this problem, we developed a sleep alarm system for drivers. This system alerts the user when he falls asleep at the wheel, thus the main concern is preventing accidents and saving lives. This system is useful for long- distance travelers and late-night drivers.