VTEC Prediction at Brazilian Region Using Artificial Neural Networks

W.C. Machado and E.S. Fonseca Jr.

Abstract:	In order to obtain precise GNSS positioning and navigation, ionospheric delay must be eliminated from data. Although the first order delay can be eliminated by means of the ion-free linear combination, it is not applicable to single frequency users. Thus, several methods based on dual frequency data from a GNSS network were developed to model ionospheric delay. However, these methods fail to provide ionospheric information in the absence of GNSS data and can hinder real-time positioning. An approach is presented to predict vertical total electron content (VTEC) using multiplayer percetrons (MLP) artificial neural network (ANN). The inputs of such a model were defined as being the ionospheric piercing point (IPP) positioning and the time, while the output is the VTEC. The seasonal and longer period variations of the ionosphere were taken into account by daily training the ANN. Tests were conducted over an area covering Brazil and its close vicinity in both high and low solar activity. The ANNs were trained using data from Global Ionospheric Maps (GIM) produced by the International GNSS Service (IGS) of the 3 previous days. The trained ANNs were used to compute the VTEC to the 72 subsequent hours and the resultant VTEC were compared to those of the GIM. The delay caused by the minimum and maximum RMS of the differences between the VTEC contained in GIM (VTECGIM) and the VTEC computed by the ANN (VTECANN) vary from 0.24 m to 1.79 m in GPS L1 frequency and from 0.44 m to 3.2 m in L5. Based on the RMS of the relative error, it can be concluded that the VTECGIM were from 70% to 87% mapped by the ANN approach.
Published in:	Proceedings of the 24th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS 2011) September 20 - 23, 2011 Oregon Convention Center, Portland, Oregon Portland, OR
Pages:	2552 - 2560
Cite this article:	Machado, W.C., Fonseca, E.S., Jr., "VTEC Prediction at Brazilian Region Using Artificial Neural Networks," Proceedings of the 24th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS 2011), Portland, OR, September 2011, pp. 2552-2560.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In