Improving One-Day-Ahead Forecasting of Low-Latitude Amplitude Scintillation Using an Upsampling-Enhanced LSTM

dc.contributor.authorMuangkammuen, Patinya
dc.contributor.authorSuthisopapan, Puripong
dc.contributor.authorTongkasem, Napat
dc.contributor.authorSupnithi, Pornchai
dc.contributor.authorKruesubthaworn, Anan
dc.contributor.authorKlenzing, Jeff
dc.contributor.authorSiritaratiwat, Apirat
dc.date.accessioned2026-08-06T10:54:05Z
dc.date.available2026-08-06T10:54:05Z
dc.date.issued2026-01-01
dc.description.abstractThe scintillation in radio wave propagation, particularly in regions near the magnetic equator, is found to be introduced by the ionospheric irregularities causing unsatisfactory performance in satellite-based applications. In order to mitigate this effect, we design a long short-term memory (LSTM) model to forecast amplitude scintillation at 1-min resolution. In addition, the upsampling-based feature preprocessing is introduced to improve forecasting performance, especially for short-term severe scintillation events. In terms of R$^{2}$, which is a popular forecast evaluation metric, our proposed model exhibits about 20% improvement over the same LSTM model without upsampling. Furthermore, although existing studies achieve good forecasting accuracy up to 4 h ahead, the proposed model sets a benchmark with one-day-ahead forecasting, but at the cost of longer training time due to upsampling.
dc.identifier.citationIEEE Transactions on Aerospace and Electronic Systems, 62, 6087-6107, 2026
dc.identifier.doi10.1109/TAES.2026.3662676
dc.identifier.issn00189251
dc.identifier.other2-s2.0-105029952533
dc.identifier.urihttps://dspace.kmitl.ac.th/handle/123456789/17726
dc.sourceIEEE Transactions on Aerospace and Electronic Systems
dc.subjectFeature preprocessing
dc.subjectionospheric scintillation
dc.subjectlong short-term memory (LSTM) network
dc.subjectmachine learning
dc.subjectpredictive models
dc.subjecttime-series forecasting
dc.titleImproving One-Day-Ahead Forecasting of Low-Latitude Amplitude Scintillation Using an Upsampling-Enhanced LSTM
dc.typeArticle

Files

Collections