accurate prediction of dgps correction using neural network trained by imperialistic competition algorithm

this paper presents an accurate differential global positioning system (dgps) using multi-layered neural networks (nns) based on the back propagation (bp) and imperialistic competition algorithm (ica) in order to predict the dgps corrections for accurate positioning. simulation results allowed us to optimize the nn performance in term of residual mean square error. we compare results obtained by the nn technique with bp and ica. results show a good improvement obtained by the application of the nn trained by the ica. the experimental results on measurement data demonstrate that the prediction total rms error using nn trained by the ica learning algorithm are 0.8273 and 0.7143 m, before and after selective availability, respectively.