A New Technique to TEC Regional Modeling using a Neural Network

Rodrigo F. Leandro

Abstract:	In this paper we present a new technique of regional modeling of TEC (Total Electron Content), using a Neural Network model. This new model has the capability to predict TEC values derived from a GPS tracking network. Preliminary tests and respective results are shown. One of the main sources of errors of GPS measurements is the ionosphere refraction. As a dispersive medium, the ionosphere allows its influence to be computed by using dual frequency receivers. The use of two frequencies allow estimating the influence of ionosphere on GPS signal by the computation of TEC values, which have a direct relationship with the magnitude of the delay caused by the ionosphere. In the case of single frequency receivers it is necessary to use models that tell us how large the ionospheric refraction is. Such is the case of which the GPS broadcast message carries parameters of the Klobuchar model. One other alternative to single frequency users is to create a regional model based on a network of dual frequency receivers. In this case, the regional behaviour of ionosphere is modeled in a way that it is possible to estimate the TEC values inside or near this region. This regional model can be based on polynomials, for example. We have investigated a Neural Network-based model to the computation of regional TEC. The advantage from the use of this Neural Network model is that with the same model we can predict values for a station either within or outside the network, due to the adaptation capability of neural networks training process, that is an iterative adjust of the synaptic weights in function of residuals, using the training parameters. We have used data from the permanent GPS tracking network in Brazil (RBMC). We have tested the accuracy of the new model at all stations. To perform the tests TEC values were computed for each station of the network, except for a test station. After that the training parameters data set for the test station was formed, based on the TEC values of all other stations of the GPS network. The Neural Network was trained with these parameters, and tested by computing the TEC for the test station. This assessment was carried out several times, one for each station of the network. Preliminary assessment of results using our new technique shows a capability of retrieving around 85 % of TEC values for all stations. This means that we can correct the ionospheric delay at the same amount, due the direct relationship between both TEC and ionospheric delay.
Published in:	Proceedings of the 17th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS 2004) September 21 - 24, 2004 Long Beach Convention Center Long Beach, CA
Pages:	366 - 374
Cite this article:	Leandro, Rodrigo F., "A New Technique to TEC Regional Modeling using a Neural Network," Proceedings of the 17th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS 2004), Long Beach, CA, September 2004, pp. 366-374.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In