Global Ionospheric Total Electron Content Forecasting Model Using Temporal Convolutional Network

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

The ionosphere is a crucial component of Earth's atmosphere, with significant impacts on radio transmission, satellite navigation, radar surveillance, and other applications. Total Electron Content (TEC) in the ionosphere is one of the vital parameters used to describe ionospheric activity and characteristics. Due to limited observational methods, we often need the TEC predictions based on historical TEC data. In this study, we employ the Temporal Convolutional Network (TCN) architecture to construct a global ionospheric TEC prediction model. We compare the TEC predictions from the TCN model under different ionospheric activity conditions with those from time-series models and deep learning recurrent structure models. The research findings indicate that the TCN-based model offers higher prediction accuracy and greater stability for global ionospheric TEC forecasts.

Original languageEnglish
Title of host publicationProceedings - 2023 Cross Strait Radio Science and Wireless Technology Conference, CSRSWTC 2023
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9798350358971
DOIs
StatePublished - 2023
Event2023 Cross Strait Radio Science and Wireless Technology Conference, CSRSWTC 2023 - Guilin, China
Duration: 10 Nov 202313 Nov 2023

Publication series

NameProceedings - 2023 Cross Strait Radio Science and Wireless Technology Conference, CSRSWTC 2023

Conference

Conference2023 Cross Strait Radio Science and Wireless Technology Conference, CSRSWTC 2023
Country/TerritoryChina
CityGuilin
Period10/11/2313/11/23

Keywords

  • Temporal Convolutional Network
  • Total Electron Content
  • ionosphere
  • predictions

Fingerprint

Dive into the research topics of 'Global Ionospheric Total Electron Content Forecasting Model Using Temporal Convolutional Network'. Together they form a unique fingerprint.

Cite this