Traffic Accident Forecast Research Articles

LLM Multimodal Traffic Accident Forecasting.

With the rise in traffic congestion in urban centers, predicting accidents has become paramount for city planning and public safety. This work comprehensively studied the efficacy of modern deep learning (DL) methods in forecasting traffic accidents and enhancing Level-4 and Level-5 (L-4 and L-5) driving assistants with actionable visual and language cues. Using a rich dataset detailing accident occurrences, we juxtaposed the Transformer model against traditional time series models like ARIMA and the more recent Prophet model. Additionally, through detailed analysis, we delved deep into feature importance using principal component analysis (PCA) loadings, uncovering key factors contributing to accidents. We introduce the idea of using real-time interventions with large language models (LLMs) in autonomous driving with the use of lightweight compact LLMs like LLaMA-2 and Zephyr-7b-α. Our exploration extends to the realm of multimodality, through the use of Large Language-and-Vision Assistant (LLaVA)-a bridge between visual and linguistic cues by means of a Visual Language Model (VLM)-in conjunction with deep probabilistic reasoning, enhancing the real-time responsiveness of autonomous driving systems. In this study, we elucidate the advantages of employing large multimodal models within DL and deep probabilistic programming for enhancing the performance and usability of time series forecasting and feature weight importance, particularly in a self-driving scenario. This work paves the way for safer, smarter cities, underpinned by data-driven decision making.

Study on Traffic Accident Forecast of Urban Excess Tunnel Considering Missing Data Filling

Urban highway tunnels are frequent accident locations, and predicting and analyzing road conditions after accidents to avoid traffic congestion is a key measure for tunnel traffic operation management. In this paper, 200 traffic accident data from the YingTian Street Tunnel in Nanjing city were analyzed and encoded to extract the main factors affecting tunnel traffic conditions from three aspects: time, traffic flow, and tunnel environment. Next, graph convolution long short-term memory networks were used to predict and fill in missing traffic flow data. Finally, seven independent variables selected by Pearson correlation analysis were input into the constructed BP neural network and random forest model to predict tunnel traffic conditions during accidents and accident duration. Experimental results show that the accuracy of random forest and BP neural networks in predicting traffic flow is 83.39% and 82.94%, respectively, and that the absolute error of the two models in predicting accident duration is 75% and 60% within 25 min, respectively. Both models perform well in predicting traffic conditions, and the random forest models perform better in terms of robustness and generalization in predicting crash duration. The experimental results have important implications for tunnel operation management during accidents.

Impact of the COVID-19 pandemic on road traffic accident forecasting in Poland and Slovakia

Abstract The COVID-19 pandemic significantly affected the performance of the transport sector and its overall intensity. Reducing mobility has a major impact on road traffic accidents. The aim of this study is to forecast the number of road traffic accidents in Poland and Slovakia and to assess how the COVID-19 pandemic affected its trend. For this purpose, data for Poland and Slovakia in the selected relevant period were analyzed. Based on actual data from the past, a forecast was made for the future considering two scenarios – one where there is no effect of pandemic, and another with effect of pandemic. Forecasting the number of accidents in Poland was carried out using selected time series models related to linear trend (Holt and Winters method) and the exponential model. In the case of Slovakia, the model without trend and the exponential model were used to forecast the number of traffic accidents. The results of the research show that the pandemic caused a decrease in the number of traffic accidents in Poland by 31% and in Slovakia by 33%. This is a significant decline, but it is linearly dependent on restrictive measures that affect the mobility of the population. A similar trend can therefore be expected on a European scale.

GSNet: Learning Spatial-Temporal Correlations from Geographical and Semantic Aspects for Traffic Accident Risk Forecasting

Traffic accident forecasting is of great importance to urban public safety, emergency treatment, and construction planning. However, it is very challenging since traffic accidents are affected by multiple factors, and have multi-scale dependencies on both spatial and temporal dimensional features. Meanwhile, traffic accidents are rare events, which leads to the zero-inflated issue. Existing traffic accident forecasting methods cannot deal with all above problems simultaneously. In this paper, we propose a novel model, named GSNet, to learn the spatial-temporal correlations from geographical and semantic aspects for traffic accident risk forecasting. In the model, a Spatial-Temporal Geographical Module is designed to capture the geographical spatial-temporal correlations among regions, while a Spatial-Temporal Semantic Module is proposed to model the semantic spatial-temporal correlations among regions. In addition, a weighted loss function is designed to solve the zero-inflated issue. Extensive experiments on two real-world datasets demonstrate the superiority of GSNet against the state-of-the-art baseline methods.

Application of big data optimized clustering algorithm in cloud computing environment in traffic accident forecast

As the usage rate of cars is getting higher and higher, the injuries and losses caused by traffic accidents are also getting bigger and bigger. If some traffic accidents can be predicted, then such losses can be greatly solved. Although there are abundant research results on intelligent transportation, there are not many research results on how to predict traffic accidents. For this issue, the main aim of this paper is to propose a continuous non-convex optimization of the K-means algorithm in order to solve the model problem in the traffic prediction process. First, this paper uses clustering algorithm for feature analysis and big data for the establishment of simulation model in cloud environment. Through this paper an equivalent model, using matrix optimization theory to analyze and process K-means problem, and design efficient and theoretically guaranteed algorithms for big data. By simulating the traffic situation in Shanghai city within three years, the outcomes display that the model endorsed in the given paper can predict traffic accidents at a rate of 93.88% and the accuracy rate of traffic accident processing time is 78%, which fully illustrates the effectiveness of the model established in this paper.

RiskOracle: A Minute-Level Citywide Traffic Accident Forecasting Framework

Real-time traffic accident forecasting is increasingly important for public safety and urban management (e.g., real-time safe route planning and emergency response deployment). Previous works on accident forecasting are often performed on hour levels, utilizing existed neural networks with static region-wise correlations taken into account. However, it is still challenging when the granularity of forecasting step improves as the highly dynamic nature of road network and inherent rareness of accident records in one training sample, which leads to biased results and zero-inflated issue. In this work, we propose a novel framework RiskOracle, to improve the prediction granularity to minute levels. Specifically, we first transform the zero-risk values in labels to fit the training network. Then, we propose the Differential Time-varying Graph neural network (DTGN) to capture the immediate changes of traffic status and dynamic inter-subregion correlations. Furthermore, we adopt multi-task and region selection schemes to highlight citywide most-likely accident subregions, bridging the gap between biased risk values and sporadic accident distribution. Extensive experiments on two real-world datasets demonstrate the effectiveness and scalability of our RiskOracle framework.

Forecast of Road Traffic Accidents grounded on Rolling optimization Grey Markov Model

This paper recommends the rolling optimization strategy based on the initial data of road traffic accidents, and builds the rolling optimization-grey Markov dynamic prediction model, which can effectively resolve the matter that the precision of accident forecast is influenced by the time benefit of the predicted data. In order to predict the development tendency of road traffic accidents and further improve the prediction precision of random time series, this paper uses Markov chain theory to probe into the transition law between different states. The case study shows that this measure has good forecast precision and practicability in a certain period of time, and can offer Reference for road traffic accident forecast and traffic safety warning.

Road Traffic Accident Scenario, Pattern and Forecasting in Bangladesh

The main aim of this research work is to be aware of the road traffic accident scenario, injurious effects and pattern in Bangladesh. Moreover we are interested to forecast the magnitude of road traffic accidents for the future so that decision makers can make appropriate decision for precaution. This study also provides an assessment of road traffic accidents in Bangladesh and its impact based on data collected for the period of 1971 to 2017. In this study we have tried to pick up the main reasons of road accidents and to observe the tremendous situation. The study observed that the general trends of road traffic accident (RTA), deaths and injuries reveal that the number of RTA, deaths and injuries increased gradually with little fluctuations form 1971 to 2007 and after 2007 there is a slow decreasing trend. Although the number of RTA and deaths observed decreasing trend in recent years, the ratio of number of deaths to number of accident increased significantly. The rate of register vehicles per 10,000 people increased moderately throughout the period but a sharp increment is exhibited from 2009. Highest percentage of RTA (34%) and deaths is due to RTA (32%) in Dhaka division while the lowest percentage of RTA (4%) in Barisal and Sylhet divisions and deaths is due to RTA (3%) in Barisal division. It is noticed that the maximum number of injuries occurred between ages 21 and 30 while the maximum number of deaths occurred between ages 11 and 30. Most of the RTA and deaths due to RTA are caused by run over by vehicles and head to head collision. The severity of occurring road accident and number of deaths are higher during the festive periods because of involving higher frequency of traveling than usual. The time plot shows that the graph maintains a decreasing movement from 2012 to 2015 but increases from 2015 to 2017. In the research an additive time series model approach is applied. It included the estimation of trend, seasonal variation and random variation using triple exponential smoothing method. We performed forecasting of RTA eliminating seasonal impact for the next three consecutive years (2018-2020) with 95% confidence interval using Holt-Winters exponential technique.

Railway Traffic Accident Forecast Based on an Optimized Deep Auto-encoder

Safety is the key point of railway transportation, and railway traffic accident prediction is the main content of safety management. There are complex nonlinear relationships between an accident and its relevant indexes. For this reason, triangular gray relational analysis (TGRA) is used for obtaining the indexes related to the accident and the deep auto-encoder (DAE) for finding out the complex relationships between them and then predicting the accident. In addition, a nonlinear weight changing particle swarm optimization algorithm, which has better convergence and global searching ability, is proposed to obtain better DAE structure and parameters, including the number of hidden layers, the number of neurons at each hidden layer and learning rates. The model was used to forecast railway traffic accidents at Shenyang Railway Bureau, Guangzhou Railway Corporation, and Nanchang Railway Bureau. The results of the experiments show that the proposed model achieves the best performance for predicting railway traffic accidents.

The Comparison on for Traffic Accident Forecasting in Long and Short Period Based on Multi-layer Recursive Forecasting Method

The multi-layer recursive forecasting method is used on road traffic accident forecasting in this study. We suggested the factors as density of population, GDP per capita, highway passenger transport, highway freight volume, highway mileage, density of road network, amount of vehicle, amount of cars per capita and environmental factors is selected by MATLAB in forecasting model. The model including autoregression item and the model including autoregression item and environmental factors is proposed. By changing the forecasting period, the forecasting results in years and months are acquired and analyzed. We conclude that the forecasting accuracy in short period is higher than in long period by comparing with results of long and short period.

Study on Urban Emergency Event Forecast - Taking Urban Traffic Accident in China as Example

In China, with the accelerated urbanization and the rapid growth of urban population, potential urban emergency risk factors gradually increased. Therefore, how to forecast the urban emergency events seems to be important. This study elaborated the mechanism of catastrophe theory and grey forecast method, established the grey cusp catastrophe model according to further research. An empirical analysis on Chinese urban traffic accident forecast was made by this model through four indicators: the number of occurrences, the number of deaths, the number of people injured and direct economic losses. By operating the model with data from 2002 to 2006, the accuracy is verified. Based on data from 2002 to 2011, the model result shows that the next traffic accident occurrence catastrophe will occur in 2015-2016, the next death number catastrophe will occur in 2012-2013, the next injured number catastrophe will occur in 2018-2019, the next direct economic loss catastrophe will occur in 2014-2015.

The use of Grey System Theory in predicting the road traffic accident in Fars province in Iran

Traffic accidents have become a more and more important factor that restrict the development of economy and threaten the safety of human beings. Considering the complexity and uncertainty of the influencing factors on traffic accidents, traffic accident forecasting can be regarded as a grey system with unknown and known information, so be analyzed by grey system theory. Grey models require only a limited amount of data to estimate the behavior of unknown systems. In this paper, first, the original predicted values of road traffic accidents are separately obtained by the GM (1,1) model, the Verhulst model and the DGM(2,1) model. The results of these models on predicting road traffic accident show that the forecasting accuracy of the GM(1,1) is higher than the Verhulst model and the DGM(2,1) model. Then, the GM(1,1) model is applied to predict road traffic accident in Fars province.

Application of Grey Model(1,1) in Road Traffic Accident Forecast

Road traffic accident forecast is a complex stochastic process. Based on the statistics of road traffic accident, Grey Model (1, 1)( short for GM(1, 1)) was applied to forecast the future number of road traffic accident in this paper. GM(1, 1) was established according to the time-series. GM(1, 1) is usually applied for such stochastic unconfirmed problem as road traffic accident forecast. Base on MATLAB software, the forecast value of road traffic accident was given. The precision of the test results show that the model is accurate and the forecast results are reliable.