Increasing of DGPS accuracy using recurrent neural networks

Differential GPS (DGPS) is used for improving of accuracy in GPS position and velocity estimations. Measurements in DGPS can be obtained as real time with a hi-h level of accuracy. This paper presents a Recurrent Neural Network (RNN) in DGPS to enhance of the position estimation. The proposed algorithm in DGPS system is implemented by a low cost commercial Coarse/Acquisition (C/A) code GPS module. The used RNN reduces errors between received position and real position. The experimental tests results with real data are stated and discussed in this paper. The results show that position components RMS errors are less than 0.5 meter after of RNNs prediction. Also, positioning accuracy by using of RNNs prediction is independence of Selective Availability error (S/A).