A complex network approach for the model of vehicle emission propagation and intelligently mine the interaction rules

With the increasing number of motor vehicles, exhaust emission has become a major source of urban pollution. Most studies are limited to the prediction of pollutant concentration, which cannot clearly indicate the change of pollution emissions and regional relationship. In this paper, we propose an...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Journal of intelligent & fuzzy systems 2022-01, Vol.42 (6), p.5375
Hauptverfasser: Zhang, Lei, Pan, Jiaxing, Xia, Pengfei, Chuyuan Wei, Jing, Changfeng, Guo, Maozu, Guo, Quansheng
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:With the increasing number of motor vehicles, exhaust emission has become a major source of urban pollution. Most studies are limited to the prediction of pollutant concentration, which cannot clearly indicate the change of pollution emissions and regional relationship. In this paper, we propose an emission propagation model of vehicle source pollution based on complex network in order to intelligently mine the interaction and propagation rules hidden behind dynamic spatiotemporal data. First, aiming at the problems of low resolution and insufficient data volume of vehicle emission data, a high-resolution pollution emission data is generated based on the COPERT (Computer Program to Calculate Emissions from Road Transport). For study the influence of causality between regions, a propagation model is designed based on the convergent cross mapping method to transform the emission time series into a complex network. In addition, we propose a novel key node mining algorithm using hybrid local and global information to identify areas of heavy pollution. Experimental results on real datasets demonstrate that the spread of pollution follows certain rules and is also affected by regional influences. Moreover, the proposed algorithm is superior to the state-of-the-art methods.
ISSN:1064-1246
1875-8967
DOI:10.3233/JIFS-211921