A long short-term memory-based framework for crash detection on freeways with traffic data of different temporal resolutions.

Traffic crash detection is a major component of intelligent transportation systems. It can explore inner relationships between traffic conditions and crash risk, prevent potential crashes, and improve road safety. However, there exist some limitations in current studies on crash detection: (1) The commonly used machine learning methods cannot simulate the evolving transitions of traffic conditions before crash occurrences; (2) Current models collected traffic data of only one temporal resolution, which cannot fully represent traffic trends in different time intervals. Therefore, this study proposes a Long short-term memory (LSTM) based framework considering traffic data of different temporal resolutions (LSTMDTR) for crash detection. LSTM is an effective deep learning method to capture the long-term dependency and dynamic transitions of pre-crash conditions. Three LSTM networks considering traffic data of different temporal resolutions are constructed, which can comprehensively indicate traffic variations in different time intervals. A fully-connected layer is used to combine the outputs of three LSTM networks, and a dropout layer is used to reduce overfitting and improve prediction performance. The LSTMDTR model is implemented on datasets of I880-N and I805-N in California, America. The results indicate that the LSTMDTR model can obtain satisfactory performance on crash detection, with the highest crash accuracy of 70.43 %. LSTMDTR models constructed on one freeway can be transferred to other similar freeways, with 65.12 % of crash accuracy on transferability. Compared with machine learning methods and LSTM models with one or two temporal resolutions, the LSTMDTR model has been validated to perform better on crash detection and transferability. A proper number of neurons in the LSTMDTR model should be determined in real applications considering acceptable detection performance and computation time. The dropout technique can reduce overfitting and improve the generalization ability of the LSTMDTR model, increasing crash accuracy from 64.49 % to 70.43 %.

[1]  Mohamed Abdel-Aty,et al.  Assessment of freeway traffic parameters leading to lane-change related collisions. , 2006, Accident; analysis and prevention.

[2]  Jun Ma,et al.  Analyzing the Leading Causes of Traffic Fatalities Using XGBoost and Grid-Based Analysis: A City Management Perspective , 2019, IEEE Access.

[3]  Mohamed Abdel-Aty,et al.  Identifying crash propensity using specific traffic speed conditions. , 2005, Journal of safety research.

[4]  Mohamed Abdel-Aty,et al.  Real-time crash prediction models: State-of-the-art, design pathways and ubiquitous requirements. , 2019, Accident; analysis and prevention.

[5]  Mohamed Abdel-Aty,et al.  Approach-level real-time crash risk analysis for signalized intersections. , 2018, Accident; analysis and prevention.

[6]  Jian Sun,et al.  A dynamic Bayesian network model for real-time crash prediction using traffic speed conditions data , 2015 .

[7]  Baozhen Yao,et al.  A support vector machine with the tabu search algorithm for freeway incident detection , 2014, Int. J. Appl. Math. Comput. Sci..

[8]  Fang Yuan,et al.  INCIDENT DETECTION USING SUPPORT VECTOR MACHINES , 2003 .

[9]  Stephen G. Ritchie,et al.  Real-time hazardous traffic condition warning system: framework and evaluation , 2005, IEEE Transactions on Intelligent Transportation Systems.

[10]  Kui Yang,et al.  Crash risk analysis for Shanghai urban expressways: A Bayesian semi-parametric modeling approach. , 2016, Accident; analysis and prevention.

[11]  Sybil Derrible,et al.  Real-time accident detection: Coping with imbalanced data. , 2019, Accident; analysis and prevention.

[12]  Jianli Xiao,et al.  SVM and KNN ensemble learning for traffic incident detection , 2019, Physica A: Statistical Mechanics and its Applications.

[13]  Jun Ma,et al.  Spatiotemporal Prediction of PM2.5 Concentrations at Different Time Granularities Using IDW-BLSTM , 2019, IEEE Access.

[14]  Wei Wang,et al.  A Genetic Programming Model for Real-Time Crash Prediction on Freeways , 2013, IEEE Transactions on Intelligent Transportation Systems.

[15]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[16]  Jun Ma,et al.  A temporal-spatial interpolation and extrapolation method based on geographic Long Short-Term Memory neural network for PM2.5 , 2019, Journal of Cleaner Production.

[17]  Dan Wang,et al.  Bi-directional long short-term memory method based on attention mechanism and rolling update for short-term load forecasting , 2019, International Journal of Electrical Power & Energy Systems.

[18]  Mohamed Abdel-Aty,et al.  Predicting Freeway Crashes from Loop Detector Data by Matched Case-Control Logistic Regression , 2004 .

[19]  Weixu Wang,et al.  Using the Bayesian updating approach to improve the spatial and temporal transferability of real-time crash risk prediction models , 2014 .

[20]  Elke A. Rundensteiner,et al.  Bidirectional LSTM-CRF for Adverse Drug Event Tagging in Electronic Health Records , 2018, Medication and Adverse Drug Event Detection.

[21]  Jack Chin Pang Cheng,et al.  A non-linear case-based reasoning approach for retrieval of similar cases and selection of target credits in LEED projects , 2015 .

[22]  Mohamed Abdel-Aty,et al.  Utilizing support vector machine in real-time crash risk evaluation. , 2013, Accident; analysis and prevention.

[23]  David Helman,et al.  TRAFFIC INCIDENT MANAGEMENT , 2004 .

[24]  Moinul Hossain,et al.  A Bayesian network based framework for real-time crash prediction on the basic freeway segments of urban expressways. , 2012, Accident; analysis and prevention.

[25]  Fei-Yue Wang,et al.  Traffic Flow Prediction With Big Data: A Deep Learning Approach , 2015, IEEE Transactions on Intelligent Transportation Systems.

[26]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[27]  Moggan Motamed.  Developing a real-time freeway incident detection model using machine learning techniques , 2016 .

[28]  Jack Chin Pang Cheng,et al.  A data-driven study of important climate factors on the achievement of LEED-EB credits , 2015 .

[29]  Dongxiang Jiang,et al.  Fault diagnosis of wind turbine based on Long Short-term memory networks , 2019, Renewable Energy.

[30]  Mohamed M. Ahmed,et al.  The Viability of Using Automatic Vehicle Identification Data for Real-Time Crash Prediction , 2012, IEEE Transactions on Intelligent Transportation Systems.

[31]  Jian Sun,et al.  Use of Support Vector Machine Models for Real-Time Prediction of Crash Risk on Urban Expressways , 2014 .

[32]  Yongsheng Chen,et al.  Predicting Real-Time Crash Risk for Urban Expressways in China , 2017 .

[33]  Jun Ma,et al.  Air quality prediction at new stations using spatially transferred bi-directional long short-term memory network. , 2020, The Science of the total environment.

[34]  Chris Lee,et al.  Real-Time Crash Prediction Model for Application to Crash Prevention in Freeway Traffic , 2002 .

[35]  Jack Chin Pang Cheng,et al.  Estimation of the building energy use intensity in the urban scale by integrating GIS and big data technology , 2016 .

[36]  Mohamed Abdel-Aty,et al.  Real-time prediction of visibility related crashes , 2012 .

[37]  Shuyan Chen,et al.  Multiple Naïve Bayes Classifiers Ensemble for Traffic Incident Detection , 2014 .