DNN Based Geoid Undulation Prediction Accuracy Evaluation Using EGM08 Gravity Model

The difference between mean sea level and ellipsoid height is defined as geoid relief, which is essential information when building barometric altimeter/GPS combined systems or terrain matching systems. Therefore, an accurate calculation of the geoid undulation is required. In this paper, we propose a deep neural network method for calculating geoid undulations in real time in an embedded system. Then, using the EGM08 geoid model of order 2160, training data at intervals of 0.001 degrees were generated, and the prediction model accuracy was evaluated for 4 cases according to the number of hidden layers. As the number of hidden layers increased, the prediction accuracy increased, and it was confirmed that the calculation time also increased proportionally. © 2022 Korean Institute of Electrical Engineers. All rights reserved.