arXiv (Cornell University) · 2018 · 24 citations · 0 references
ComputingAerospace EngineeringEarth SciencesForecastingGlobal Navigation Satellite SystemsMachine LearningDeep LearningLong-term ForecastingTime Series ForecastingTime Series ModelingTime Series PredictionConvolutional Forecasting ModelsConvolutional Neural NetworksDeep Learning For ForecastingDeep Learning For Time Series ForecastingDeep Learning-based ForecastingIonospheric Total Electron ContentRecurrent Neural NetworksTotal Electron Content Modeling
The ionosphere electromagnetic activity is a major factor of the quality of\nsatellite telecommunications, Global Navigation Satellite Systems (GNSS) and\nother vital space applications. Being able to forecast globally the Total\nElectron Content (TEC) would enable a better anticipation of potential\nperformance degradations. A few studies have proposed models able to predict\nthe TEC locally, but not worldwide for most of them. Thanks to a large record\nof past TEC maps publicly available, we propose a method based on Deep Neural\nNetworks (DNN) to forecast a sequence of global TEC maps consecutive to an\ninput sequence of TEC maps, without introducing any prior knowledge other than\nEarth rotation periodicity. By combining several state-of-the-art\narchitectures, the proposed approach is competitive with previous works on TEC\nforecasting while predicting the TEC globally.\n