Prediction of Differential GPS Corrections Using AR and ARMA Models

Abstract

Global Positioning System provides user position, velocity and time anywhere on the surface of the earth. Navigation accuracy and security requirements are highly crucial for defence applications, such as, missile tracking and guidance. The accuracy of standalone GPS is not sufficient for such applications. So the accuracy can be improved with the help of Differential GPS (DGPS). In DGPS, the differential corrections are transmitted to the rovers which are in the vicinity of reference station. If there is any loss in transmission of differential corrections it may lead to navigation inaccuracy. This problem can be minimized by introducing Autoregressive (AR) and Autoregressive Moving Average (ARMA) models at the rover to predict the corrections. The Mean square error (MSE) of DGPS Correction predicted values are computed and observed that the MSE value due to ARMA model is 0.0968, 0.085 for PRN 6, 19 respectively and is better when compared to AR model.