Learning spatial patterns and temporal dependencies for traffic accident severity prediction: A deep learning approach

The article provides an analysis of the global trends in the field of city transport systems' safety. It is shown that the decrease in the safety of traffic is one of the consequences of the growth of motorization. The efficiency of measures to prevent traffic accidents is analyzed from the viewpoint of their role in the process to ensure safety and sustainability of the urban transportation system. Currently, the efforts to increase the proportion of trips made by foot or by non-motorized transport are made. That is why the problem of ensuring safety of the most vulnerable road users is relevant. One of methods to prevent traffic accidents is to identify potentially dangerous road users, whose behavior may lead to serious consequences. Technical equipment, as well as active control from the side of the services, that ensure road safety, are used to identify violators. Official statistics of the State Inspectorate for Traffic Security is used as initial information. According to the Haddon matrix, there should be 3 stages to decrease the number, as well as the severity of traffic accidents. From our point of view, the most important stage is the phase “before the traffic accident”. The statistics of traffic accidents with victims, as well as the factors that cause such accidents, are analyzed. Simulation has shown, that rational management can reduce the likelihood and the severity of road traffic accidents. Ways to improve efficiency of the above mentioned traffic accidents' prevention measures are proposed.

Traffic accidents are one of the leading causes of death and disability, as well as a significant source of economic losses for society. However, the nonlinear and heterogeneous relationships between environmental factors and traffic accidents are complex and difficult to comprehend. This study constructs an explainable spatial machine learning framework using a geographically weighted support vector machine (GW‐SVM) model to address issues of nonlinearity, spatial heterogeneity, and interpretability. Based on a large‐scale traffic accident dataset and multisource big data, this study provides both global and local explanations for the nonlinear relationships in California, USA. The study finds that (1) humidity plays a more important role in the relationship between environmental factors and traffic accident severity; (2) all environmental variables, including both natural and socioeconomic variables, exhibit nonlinear and threshold effects on traffic accidents; and (3) compared to the existing models, the GW‐SVM model performs better in predicting the severity of traffic accidents on urban roads. The results of this study are significant for reducing traffic accident risks.

Vulnerable road users, including motorcyclists, account for more than half of all road traffic fatalities worldwide. Communication actions are an essential part of prevention both in Africa and in the world. This study aimed to review the available scientific studies on the effectiveness of communication actions for the reduction of road traffic accidents among motorcyclists. We carried out an automatic search on PubMed using an equation and a manual search on Google Scholar using keywords and the bibliographic references of the studies generated by the automatic search. Using this strategy, 155 articles were identified, 16 of which were eliminated as duplicates. Thus, 139 were screened by reading the titles, abstracts, and full-texts. In the end, 13 articles met the inclusion criteria. The included articles provided evidence of the effectiveness of communication actions on the reduction of risk factors for the occurrence or severity of road traffic accidents. However, communication actions alone seem insufficient to achieve a sustainable and significant reduction in road traffic accidents, road traffic injuries, or road traffic fatalities. This confirms the need to integrate all road safety promotion strategies to act simultaneously on the various risk factors for the occurrence or severity of road traffic accidents and thereby significantly reduce their incidence and associated consequences.

The increasing complexity of urban transportation systems and the growing volume of vehicles have made traffic congestion a persistent challenge in modern cities. Efficient traffic flow prediction is essential for mitigating congestion, improving road safety, optimizing traffic signal control, and enhancing overall transportation efficiency. In recent years, artificial intelligence (AI) has emerged as a transformative tool in the field of traffic management, offering sophisticated algorithms capable of modeling, analyzing, and predicting complex traffic patterns with high accuracy. The application of AI in traffic flow prediction leverages vast amounts of real-time and historical data to generate precise forecasts, supporting data-driven decision-making by urban planners and traffic control authorities. The prediction of traffic flow involves analyzing time-series data that exhibit nonlinear, dynamic, and often stochastic behavior. Traditional statistical models, such as autoregressive integrated moving average (ARIMA), have proven to be limited in handling the high dimensionality and variability inherent in traffic systems. In contrast, AI algorithms possess the capacity to learn and adapt from complex data inputs without the need for explicit programming, making them particularly suitable for traffic-related applications. AI algorithms used in traffic flow prediction can be broadly categorized into machine learning (ML) and deep learning (DL) approaches. Machine learning algorithms such as k-nearest neighbors (KNN), support vector machines (SVM), decision trees, and random forests have demonstrated effectiveness in short-term traffic prediction tasks. These algorithms are capable of identifying hidden patterns in traffic data and adjusting to changes in traffic behavior over time. Ensemble methods, which combine the strengths of multiple learning models, further enhance prediction accuracy and robustness. Deep learning algorithms, a subfield of AI inspired by the human brain’s neural architecture, have shown exceptional performance in capturing spatial-temporal dependencies in traffic data. Recurrent neural networks (RNNs), particularly long short-term memory (LSTM) networks and gated recurrent units (GRUs), are widely used for their ability to process sequential data and retain information over extended time intervals. Convolutional neural networks (CNNs) are employed to extract spatial features from traffic sensor data or road network imagery. Hybrid models that integrate CNNs with RNNs have achieved high levels of predictive precision by simultaneously learning spatial and temporal correlations. In addition to supervised learning methods, unsupervised and reinforcement learning techniques are also applied in traffic flow prediction. Clustering algorithms, such as k-means and DBSCAN, assist in identifying traffic patterns, while reinforcement learning models optimize adaptive traffic signal control systems by learning optimal actions through environmental interaction. This study explores the different types of AI algorithms used in traffic flow prediction, examining their theoretical foundations, structural differences, and practical applications. It aims to evaluate the comparative advantages of various algorithms in addressing the challenges of real-time traffic prediction in increasingly complex transportation networks.

Traffic incidents continue to cause a significant loss in deaths, injuries, and property damages. Reported traffic accident data contains a considerable amount of human errors, hindering the studies on traffic accidents. Several approaches have been developed to detect accidents using traffic data in real time. However, those approaches do not consider the spatiotemporal patterns inherent in traffic data, resulting in high false alarm rates. In this paper, we study the problem of traffic accident detection by considering multiple traffic speed time series collected from road network sensors. To capture the spatiotemporal impact of traffic accidents to upstream locations, we adopt Impact Interval Grouping (IIG), which compares real-time traffic speed with historical data, and generates impact intervals to determine the presence of accidents. Furthermore, we take a multivariate time series classification approach and extract three novel features to measure the severity of traffic accidents. We use real-world traffic speed and accident datasets in our empirical evaluation, and our solutions outperform state-of-the-art approaches in multivariate time series classification.

The latest developments in Advanced Driver Assistance Systems (ADAS) have greatly enhanced the comfort and safety of drivers. These technologies can identify driver abnormalities like fatigue, inattention, and impairment, which are essential for averting collisions. One of the important aspects of this technology is automated traffic accident detection and prediction, which may help in saving precious human lives. This study aims to explore critical features related to traffic accident detection and prevention. A public US traffic accident dataset was used for the aforementioned task, where various machine learning (ML) models were applied to predict traffic accidents. These ML models included Random Forest, AdaBoost, KNN, and SVM. The models were compared for their accuracies, where Random Forest was found to be the best-performing model, providing the most accurate and reliable classification of accident-related data. Owing to the black box nature of ML models, this best-fit ML model was executed with explainable AI (XAI) methods such as LIME and permutation importance to understand its decision-making for the given classification task. The unique aspect of this study is the introduction of explainable artificial intelligence which enables us to have human-interpretable awareness of how ML models operate. It provides information about the inner workings of the model and directs the improvement of feature engineering for traffic accident detection, which is more accurate and dependable. The analysis identified critical features, including sources, descriptions of weather conditions, time of day (weather timestamp, start time, end time), distance, crossing, and traffic signals, as significant predictors of the probability of an accident occurring. Future ADAS technology development is anticipated to be greatly impacted by the study’s conclusions. A model can be adjusted for different driving scenarios by identifying the most important features and comprehending their dynamics to make sure that ADAS systems are precise, reliable, and suitable for real-world circumstances.

With the increase of truck ownership year by year, traffic accidents involving heavy trucks occur frequently and the consequence is more serious. The accident is the result of the maladjustment of various factors in the “man-vehicle-environment” system. In this context, using the data of 305 traffic accidents of heavy truck in a province of China in 2020, 11 influencing factors are selected from the three aspects of people, vehicle, and environment. The scientific analysis of the factors affecting the severity of heavy truck accidents is carried out by the Logistic regression model. Mining the influence of “man-machine-environment” system disharmony on the severity of the accident. The result shows that the main factors affecting the severity of heavy truck accidents are driving on the wrong side of the road, failing to yield as required, involving pedestrians, vehicle fault and without traffic control. Based on the above five factors, countermeasures to prevent heavy truck traffic accidents are put forward in order to reduce casualties and property damage caused by accidents. It is of great theoretical significance and practical application value to improve the level of road traffic safety in China.KeywordsMan-machine-environmentHeavy goods vehicleAccident severityLogistic regression models

Predictive analysis on cellular traffic is important for the control and monitoring of wireless networks. Cellular traffic prediction is a challenging problem due to the non- stationarity and dynamic spatial-temporal correlation of the traffic. In this paper, we address the problem of accurate traffic prediction in a base station by proposing a deep neural network called RAConv. Its structure includes residual network, attention mechanism, and deep convolutional network. In the proposed architecture, a deep 3D residual convolutional network (ResConv3D) with three residual blocks are employed to learn the local spatial-temporal features. An attention-aided convolutional long short-term memory network (AConvLSTM) is then used to capture the long-term spatial-temporal dependencies. The use of the attention modules enable the network to focus on the most important spatial-temporal information. We evaluate the performance of the proposed RAConv network using a dataset provided by a Canadian wireless service provider. We consider the traffic prediction on two time scales (i.e., hourly and daily), which exhibit different spatial-temporal dependency patterns. Experimental results show that the proposed RAConv network can achieve accurate prediction under both time scales. Results also show that our proposed network provides a lower root-mean-square error (RMSE) than the conventional ConvLSTM baseline scheme.

Road accidents are a major critical problem in recent days, which cause death, severe injuries, disabilities to human lives, and substantial economic losses worldwide. A detailed analysis of the factors responsible for the accidents is required to reduce the accident rate. In this study, a descriptive analysis has been performed in-depth to identify the significant factors influencing accident severity. Machine learning techniques have been applied to predict the severity of accidents considering location, time, infrastructure, and environmental conditions-related factors that may cause road accidents. Here the findings of this study emphasize that the mediate severity of accidents have a high frequency of occurring rather than severities of very low and high risk. Moreover, the factors like infrastructure, day of the week, and weather conditions influence the severity of accidents differently. The Random Forest algorithm took the best performance with 97.2% high accuracy in predicting the severity of road traffic accidents. These results can guide relevant parties to identify dangerous situations and take the necessary actions to improve road safety by reducing accidents.

This study focused on estimating influential factors of traffic accidents and analyzing traffic accident severity of elderly and non elderly using traffic accident data. In order to reclassify elderly and non elderly traffic accident by a statistical method from entire traffic accident data, multiple discriminant analysis was applied. Also ordered logit model was applied for analyzing traffic accident severities using traffic accident severities as an independent variable and transportation facilities, road conditions and human characteristics as dependent variables. As results of the comparison between elderly and non elderly traffic accident, the traffic accident severity was affected by the age, types of traffic accidents, human characteristics and road conditions as well. Also, transportation facilities and road conditions affected to more elderly traffic accident than non elderly. Therefore, traffic accident severity would be decreased with the improvement of transportation facilities and road conditions for the elderly.

Background/Objectives: This paper presents a Residual Neural Network (ResNet) based framework tailored for structured traffic accident data, aiming to improve accident severity prediction. The proposed model leverages residual learning to effectively model intricate relationships between numerical and categorical variables, resulting in a notable increase in prediction accuracy. Methods: A comparative analysis was performed with other Deep Learning (DL) architectures, including Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM), Darknet, and Extreme Inception (Xception), showing superior performance of the proposed Resnet. Key factors influencing accident severity were identified, with Shapley Additive Explanations (SHAP) values helping to address the need for transparent and explainable Artificial Intelligence (AI) in critical decision-making areas. Results: The generalizability of the ResNet model was assessed by training it, initially, on a UK road accidents dataset and validating it on a distinct dataset from India. The model consistently demonstrated high predictive accuracy, underscoring its robustness across diverse contexts, despite regional differences. Conclusions: These results suggest that the adapted ResNet model could significantly enhance traffic safety evaluations and contribute to the formulation of more effective traffic management strategies.

Traffic accident prediction on road levels and minute levels plays an important role in optimizing public safety and improving traffic infrastructure. However, there are still some challenges in this work. Firstly, the dynamic factors (e.g. traffic flow) affecting traffic accidents make the road network have dynamic spatio‐temporal dependency, which leads to biased prediction results. Secondly, the occurrence of traffic accidents is a small probability event, which brings about zero‐inflation problem. To address aforementioned problems, the authors propose a Multi‐Attention Dynamic Graph Convolution Network with Cost Sensitive Learning approach (MADGCN). Specifically, in the spatial dimension, MADGCN calculates the attention scores of different types of dynamic factors through attention mechanism to simulate the different influence degrees of different factors, and models dynamic inter‐road spatial correlation through Graph Convolution Network (GCN). In the temporal aspect, MADGCN adaptively models the dynamic temporal correlations through self‐attention blocks. In addition, MADGCN improves the loss function based on cost‐sensitive learning strategy to increase the cost of false classification of positive samples, so as to accurately mine sparse positive samples. Experimental results on two real‐world traffic accident datasets demonstrate the superiority of MADGCN. Compared with the existing state‐of‐the‐art methods, F1‐score of MADGCN on two real‐world traffic accident datasets is improved by 11.61% and 9.15%, respectively.

The rarity and randomness of serious traffic accidents result in highly imbalanced datasets, posing challenges for prediction and interpretation. Although emerging deep learning methods effectively address class imbalance, they are often unstable and may generate unrealistic data. To address this issue, this study proposes using the K-Prototype (KP) clustering algorithm to refine synthetic data, with Gini impurity employed to filter high-quality data by Wasserstein Generative Adversarial Network with Gradient Penalty (WGAN-GP). Subsequently, we develop an evaluation framework to assess the effectiveness of the proposed method, including a systematic comparison of the similarity between synthetic data generated by different resampling techniques and real data, as well as their impact on improving accident severity classification performance. The results show that the proposed method not only maintains data diversity but also enhances feature distribution, with the synthesized data from WGAN-GP-KP outperforming all other methods in accident severity prediction models. Furthermore, the SHapley Additive exPlanations (SHAP) technique is employed to analyze the factors influencing the severity of traffic accidents. This study offers valuable insights for traffic researchers in addressing imbalanced datasets related to accident severity.

The prediction of traffic accident severity is essential for traffic safety management and control. To achieve high prediction accuracy and model interpretability, we propose a hybrid model that integrates random forest (RF) and Bayesian optimization (BO). In the proposed model, BO-RF, RF is adopted as a basic predictive model and BO is used to tune the parameters of RF. Experimental results show that BO-RF achieves higher accuracy than conventional algorithms. Moreover, BO-RF provides interpretable results by relative importance and a partial dependence plot. We can identify important influential factors for traffic accident severity by relative importance. Further, we can investigate how the influential factors affect traffic accident severity by the partial dependence plot. These results provide insights to mitigate the severity of traffic accident consequences and contribute to the sustainable development of transportation.

Road traffic accidents are a common and seemingly inevitable problem. While its occurrences rely on many unpredictable factors, this paper shows how to utilize machine learning to predict both the possibility of the accident and its severity. The datasets used were related to road accidents in several countries in a period of a few years. Some of the parameters observed were the weather conditions, sun position, speed limit, and time of the day. To predict the severity of the accident given the circumstances and road conditions, a multiclass classification model is used. Different datasets were combined to cover different situations and scenarios that happen in traffic and taking the severity of accidents in prediction. The dataset values were normalized before the training process and the training set and validated on the validation dataset. The prediction results show the correlation between used weather conditions, daylight time, and traffic accident severity.