EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Application of convolutional neural networks in spatio-temporal prediction of vehicular traffic

Cristian Camilo Montenegro Basante

SAP Engineering and Industry · 2026

DOI: 10.62486/ei2026102 ↗

Kurz gefasst

Fachartikel in der Zeitschrift SAP Engineering and Industry, 2026. Verfasst von Cristian Camilo Montenegro Basante. Die Arbeit nennt 58 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This paper systematically reviews how convolutional neural networks (CNNs) have been used to predict vehicular traffic in space and time. To this end, research published between 2021 and 2026 in different academic databases was analyzed, with the aim of identifying the most used architectures, methodological approaches, evaluation metrics, as well as their main limitations and recent trends. From this review, it is observed that CNNs stand out especially in the identification of spatial traffic patterns. Their performance improves when combined with other approaches, such as recurrent models, attention mechanisms, and hybrid architectures, allowing traffic dynamics to be captured more completely. In particular, hybrid models have managed to increase accuracy by integrating both spatial and temporal information, in addition to incorporating external variables such as weather, accidents, events, and road conditions. However, the literature also points to several challenges: adequately representing complex road networks, avoiding the loss of topological information, modeling long-term dependencies, and reducing high computational costs, especially in real-time applications. Together, these findings offer a clear view of the current state of the field and serve as a starting point for the development of future solutions oriented to smart mobility systems.

Zitieren

APA

Basante, C. C. M. (2026). Application of convolutional neural networks in spatio-temporal prediction of vehicular traffic. SAP Engineering and Industry. https://doi.org/10.62486/ei2026102

Vancouver

Basante CCM. Application of convolutional neural networks in spatio-temporal prediction of vehicular traffic. SAP Engineering and Industry. 2026. doi:10.62486/ei2026102

BibTeX
@article{Basante2026,
  author = {Basante, Cristian Camilo Montenegro},
  title = {Application of convolutional neural networks in spatio-temporal prediction of vehicular traffic},
  journal = {SAP Engineering and Industry},
  year = {2026},
  doi = {10.62486/ei2026102}
}
Kostenlos starten →