A framework for detecting and predicting highway traffic anomalies via multimodal fusion and heterogeneous graph neural networks.

This paper presents a novel framework for detecting and predicting abnormal traffic events on highways. Current traffic monitoring systems often rely on single data sources, which limits their detection accuracy and robustness in complex environments. To address these limitations, we propose a frame...

Full description

Saved in:
Bibliographic Details
Main Authors: Shaowei Sun, Mingzhou Liu
Format: Article
Language:English
Published: Public Library of Science (PLoS) 2025-01-01
Series:PLoS ONE
Online Access:https://doi.org/10.1371/journal.pone.0326313
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849424719734898688
author Shaowei Sun
Mingzhou Liu
author_facet Shaowei Sun
Mingzhou Liu
author_sort Shaowei Sun
collection DOAJ
description This paper presents a novel framework for detecting and predicting abnormal traffic events on highways. Current traffic monitoring systems often rely on single data sources, which limits their detection accuracy and robustness in complex environments. To address these limitations, we propose a framework based on multimodal deep fusion and heterogeneous graph neural networks (HGNNs), incorporating an Ensemble Contrastive Pessimistic Likelihood Estimation (CPLE) algorithm to optimize performance. The framework integrates static and dynamic traffic data, such as video images, traffic flow, vehicle speed, and tunnel weather conditions. Experimental results demonstrate that the model performs well in various scenarios, showing significant improvement in accuracy and stability over existing models like AGC-LSTM and AttentionDeepST. For instance, the proposed MHGNN-CPLE model achieves superior accuracy and F1 score in static detection tasks while maintaining high accuracy under different noise levels in dynamic detection scenarios. This study provides a significant advancement in traffic event analysis by effectively combining multimodal data and leveraging HGNNs to capture complex spatiotemporal dependencies.
format Article
id doaj-art-e709e85a99034b208fe5b72d5f430382
institution Kabale University
issn 1932-6203
language English
publishDate 2025-01-01
publisher Public Library of Science (PLoS)
record_format Article
series PLoS ONE
spelling doaj-art-e709e85a99034b208fe5b72d5f4303822025-08-20T03:30:03ZengPublic Library of Science (PLoS)PLoS ONE1932-62032025-01-01206e032631310.1371/journal.pone.0326313A framework for detecting and predicting highway traffic anomalies via multimodal fusion and heterogeneous graph neural networks.Shaowei SunMingzhou LiuThis paper presents a novel framework for detecting and predicting abnormal traffic events on highways. Current traffic monitoring systems often rely on single data sources, which limits their detection accuracy and robustness in complex environments. To address these limitations, we propose a framework based on multimodal deep fusion and heterogeneous graph neural networks (HGNNs), incorporating an Ensemble Contrastive Pessimistic Likelihood Estimation (CPLE) algorithm to optimize performance. The framework integrates static and dynamic traffic data, such as video images, traffic flow, vehicle speed, and tunnel weather conditions. Experimental results demonstrate that the model performs well in various scenarios, showing significant improvement in accuracy and stability over existing models like AGC-LSTM and AttentionDeepST. For instance, the proposed MHGNN-CPLE model achieves superior accuracy and F1 score in static detection tasks while maintaining high accuracy under different noise levels in dynamic detection scenarios. This study provides a significant advancement in traffic event analysis by effectively combining multimodal data and leveraging HGNNs to capture complex spatiotemporal dependencies.https://doi.org/10.1371/journal.pone.0326313
spellingShingle Shaowei Sun
Mingzhou Liu
A framework for detecting and predicting highway traffic anomalies via multimodal fusion and heterogeneous graph neural networks.
PLoS ONE
title A framework for detecting and predicting highway traffic anomalies via multimodal fusion and heterogeneous graph neural networks.
title_full A framework for detecting and predicting highway traffic anomalies via multimodal fusion and heterogeneous graph neural networks.
title_fullStr A framework for detecting and predicting highway traffic anomalies via multimodal fusion and heterogeneous graph neural networks.
title_full_unstemmed A framework for detecting and predicting highway traffic anomalies via multimodal fusion and heterogeneous graph neural networks.
title_short A framework for detecting and predicting highway traffic anomalies via multimodal fusion and heterogeneous graph neural networks.
title_sort framework for detecting and predicting highway traffic anomalies via multimodal fusion and heterogeneous graph neural networks
url https://doi.org/10.1371/journal.pone.0326313
work_keys_str_mv AT shaoweisun aframeworkfordetectingandpredictinghighwaytrafficanomaliesviamultimodalfusionandheterogeneousgraphneuralnetworks
AT mingzhouliu aframeworkfordetectingandpredictinghighwaytrafficanomaliesviamultimodalfusionandheterogeneousgraphneuralnetworks
AT shaoweisun frameworkfordetectingandpredictinghighwaytrafficanomaliesviamultimodalfusionandheterogeneousgraphneuralnetworks
AT mingzhouliu frameworkfordetectingandpredictinghighwaytrafficanomaliesviamultimodalfusionandheterogeneousgraphneuralnetworks