A Data-Driven Deep Learning Framework for Prediction of Traffic Crashes at Road Intersections
Traffic crash prediction (TCP) is a fundamental problem for intelligent transportation systems in smart cities. Improving the accuracy of traffic crash prediction is important for road safety and effective traffic management. Owing to recent advances in artificial neural networks, several new deep-l...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-01-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/15/2/752 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832589212197584896 |
|---|---|
| author | Mengxiang Wang Wang-Chien Lee Na Liu Qiang Fu Fujun Wan Ge Yu |
| author_facet | Mengxiang Wang Wang-Chien Lee Na Liu Qiang Fu Fujun Wan Ge Yu |
| author_sort | Mengxiang Wang |
| collection | DOAJ |
| description | Traffic crash prediction (TCP) is a fundamental problem for intelligent transportation systems in smart cities. Improving the accuracy of traffic crash prediction is important for road safety and effective traffic management. Owing to recent advances in artificial neural networks, several new deep-learning models have been proposed for TCP. However, these works mainly focus on accidents in regions, which are typically pre-determined using a grid map. We argue that TCP for roads, especially for crashes at or near road intersections which account for more than 50% of the fatal or injury crashes based on the Federal Highway Administration, has a significant practical and research value and thus deserves more research. In this paper, we formulate <i>T</i>CP at Road Intersections as a classification problem and propose a three-phase data-driven deep learning model, called <i>Road Intersection Traffic Crash Prediction</i> (<i>RoadInTCP</i>), to predict traffic crashes at intersections by exploiting publicly available heterogeneous big data. In Phase I we extract discriminative latent features called <i>topological-relational features (tr-features)</i>, of intersections using a neural network model by exploiting topological information of the road network and various relationships amongst nearby intersections. In Phase II, in addition to <i>tr-features</i> which capture some inherent properties of the road network, we also explore additional thematic information in terms of environmental, traffic, weather, risk, and calendar features associated with intersections. In order to incorporate the potential correlation in nearby intersections, we utilize a Graph Convolution Network (GCN) to aggregate features from neighboring intersections based on a message-passing paradigm for TCP. While Phase II serves well as a TCP model, we further explore the signals embedded in the sequential feature changes over time for TCP in Phase III, by exploring RNN or 1DCNN which have known success on sequential data. Additionally, to address the serious issues of imbalanced classes in TCP and large-scale heterogeneous big data, we propose an effective data sampling approach in data preparation to facilitate model training. We evaluate the proposed <i>RoadInTCP</i> model via extensive experiments on a real-world New York City traffic dataset. The experimental results show that the proposed <i>RoadInTCP</i> robustly outperforms existing methods. |
| format | Article |
| id | doaj-art-b1591a7f2af24838971c682bfcfcdfb6 |
| institution | Kabale University |
| issn | 2076-3417 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj-art-b1591a7f2af24838971c682bfcfcdfb62025-01-24T13:20:43ZengMDPI AGApplied Sciences2076-34172025-01-0115275210.3390/app15020752A Data-Driven Deep Learning Framework for Prediction of Traffic Crashes at Road IntersectionsMengxiang Wang0Wang-Chien Lee1Na Liu2Qiang Fu3Fujun Wan4Ge Yu5China National Institute of Standardization, Beijing 100088, ChinaDepartment of Computer Science and Engineering, The Pennsylvania State University, University Park, PA 16802, USAChina National Institute of Standardization, Beijing 100088, ChinaChina National Institute of Standardization, Beijing 100088, ChinaChina National Institute of Standardization, Beijing 100088, ChinaSchool of Computer Science and Engineering, Northeastern University, Shenyang 110136, ChinaTraffic crash prediction (TCP) is a fundamental problem for intelligent transportation systems in smart cities. Improving the accuracy of traffic crash prediction is important for road safety and effective traffic management. Owing to recent advances in artificial neural networks, several new deep-learning models have been proposed for TCP. However, these works mainly focus on accidents in regions, which are typically pre-determined using a grid map. We argue that TCP for roads, especially for crashes at or near road intersections which account for more than 50% of the fatal or injury crashes based on the Federal Highway Administration, has a significant practical and research value and thus deserves more research. In this paper, we formulate <i>T</i>CP at Road Intersections as a classification problem and propose a three-phase data-driven deep learning model, called <i>Road Intersection Traffic Crash Prediction</i> (<i>RoadInTCP</i>), to predict traffic crashes at intersections by exploiting publicly available heterogeneous big data. In Phase I we extract discriminative latent features called <i>topological-relational features (tr-features)</i>, of intersections using a neural network model by exploiting topological information of the road network and various relationships amongst nearby intersections. In Phase II, in addition to <i>tr-features</i> which capture some inherent properties of the road network, we also explore additional thematic information in terms of environmental, traffic, weather, risk, and calendar features associated with intersections. In order to incorporate the potential correlation in nearby intersections, we utilize a Graph Convolution Network (GCN) to aggregate features from neighboring intersections based on a message-passing paradigm for TCP. While Phase II serves well as a TCP model, we further explore the signals embedded in the sequential feature changes over time for TCP in Phase III, by exploring RNN or 1DCNN which have known success on sequential data. Additionally, to address the serious issues of imbalanced classes in TCP and large-scale heterogeneous big data, we propose an effective data sampling approach in data preparation to facilitate model training. We evaluate the proposed <i>RoadInTCP</i> model via extensive experiments on a real-world New York City traffic dataset. The experimental results show that the proposed <i>RoadInTCP</i> robustly outperforms existing methods.https://www.mdpi.com/2076-3417/15/2/752machine learningartificial intelligenceneural networkdata miningroad intersectiontraffic crash prediction |
| spellingShingle | Mengxiang Wang Wang-Chien Lee Na Liu Qiang Fu Fujun Wan Ge Yu A Data-Driven Deep Learning Framework for Prediction of Traffic Crashes at Road Intersections Applied Sciences machine learning artificial intelligence neural network data mining road intersection traffic crash prediction |
| title | A Data-Driven Deep Learning Framework for Prediction of Traffic Crashes at Road Intersections |
| title_full | A Data-Driven Deep Learning Framework for Prediction of Traffic Crashes at Road Intersections |
| title_fullStr | A Data-Driven Deep Learning Framework for Prediction of Traffic Crashes at Road Intersections |
| title_full_unstemmed | A Data-Driven Deep Learning Framework for Prediction of Traffic Crashes at Road Intersections |
| title_short | A Data-Driven Deep Learning Framework for Prediction of Traffic Crashes at Road Intersections |
| title_sort | data driven deep learning framework for prediction of traffic crashes at road intersections |
| topic | machine learning artificial intelligence neural network data mining road intersection traffic crash prediction |
| url | https://www.mdpi.com/2076-3417/15/2/752 |
| work_keys_str_mv | AT mengxiangwang adatadrivendeeplearningframeworkforpredictionoftrafficcrashesatroadintersections AT wangchienlee adatadrivendeeplearningframeworkforpredictionoftrafficcrashesatroadintersections AT naliu adatadrivendeeplearningframeworkforpredictionoftrafficcrashesatroadintersections AT qiangfu adatadrivendeeplearningframeworkforpredictionoftrafficcrashesatroadintersections AT fujunwan adatadrivendeeplearningframeworkforpredictionoftrafficcrashesatroadintersections AT geyu adatadrivendeeplearningframeworkforpredictionoftrafficcrashesatroadintersections AT mengxiangwang datadrivendeeplearningframeworkforpredictionoftrafficcrashesatroadintersections AT wangchienlee datadrivendeeplearningframeworkforpredictionoftrafficcrashesatroadintersections AT naliu datadrivendeeplearningframeworkforpredictionoftrafficcrashesatroadintersections AT qiangfu datadrivendeeplearningframeworkforpredictionoftrafficcrashesatroadintersections AT fujunwan datadrivendeeplearningframeworkforpredictionoftrafficcrashesatroadintersections AT geyu datadrivendeeplearningframeworkforpredictionoftrafficcrashesatroadintersections |