Abstract
Geometric dilution of precision (GDOP) is a powerful, simple and widely used measure for assessing the effectiveness of potential measurements to specify the precision and accuracy of the data received from global positioning system (GPS) satellites. The most correct method to classify or approximate the GPS GDOP is to use inverse matrix on all the combinations and choosing the lowest one, but inversing a matrix puts a lot of computational burden on the navigator's processor. This approach however is a time-consuming task. To overcome the problem, basic back propagation neural network (BPNN) was used. Since the BPNN is too slow for practical problems, including the GPS GDOP classification, in this paper several methods, namely, resilient back propagation (RBP) to train a NN, naive Bayes classifier, Fisher's linear discriminant (FLD) and k-nearest neighbor (KNN) for classification of the GPS GDOP are proposed. Simulation results show that these methods are much more efficient to classify the GPS GDOP data than previous methods. © 2013 IEEE.