The Satellite Selection Algorithm of GNSS Based on Neural Network

Abstract

With the development of Global Navigation Satellite System (GNSS), there is an increase in the number of visible satellite in navigation and positioning. Calculating all visible satellite not only makes the hardware design of the receiver harder, but also can’t guarantee to improve the accuracy of the positioning results, so it is significant to take a certain satellite selection algorithm to the visible satellites optimization. The traditional optimal satellite selection algorithm is complicated, and it also influences the real-time performance of the navigation and positioning. The subsequent improved satellite selection algorithm has effectively enhanced in robustness, accuracy and real-time. The satellite selection algorithm by using Generalized Regression Neural Network (GRNN) to predict GDOP is proposed. The simulation results showed that the algorithm has improved in robustness, accuracy and real-time, and it has certain exploring value for the GNSS satellite selection algorithm with machine learning.