Single-frequency GPS receivers ionospheric time-delay approximation using radial basis function neural network

Ionospheric errors are the most influential source of the GPS positioning errors. In an attempt to at least partially improve the GPS positioning process, a global ionospheric correction model is introduced within the standard GPS positioning service. Referred to as the Klobuchar model after its inventor, this model appeared as a compromise between computation complexity and correction accuracy. Neural Networks (NNs) are ideal tool for the approximation of single-frequency GPS receivers ionospheric time-delay behaviour, which has nature highly non-linear. A major advantage of using NNs for the approximation of single-frequency GPS receivers ionospheric time-delay behaviour over analytical methods is that no previous knowledge of the nature of the non-linear relationships is required. This paper presents the use of NN modeling to approximate single-frequency GPS receivers ionospheric time-delay to reduce GPS signal propagation error. The ionospheric time-delay is approximated utilizing the Radial Basis Function (RBF) NN approach. The NN-based approach reduces the computational burden respect to Klobuchar model. The method employed here is applicable on the L1 GPS receivers.