Deep Learning Method for Evaporation Duct Inversion Based on GPS Signal

Accurate evaporation duct prediction is one of the critical technologies for realizing over-the-horizon impact marine communication, ship radar, and other systems. Using GPS signals to invert ducts provides more benefits in terms method realization ease. In order from GPS-received power data, a deep learning technique based on Bayesian optimization proposed increase accuracy ducts. The propagation mechanism signal explored. estimated using two-parameter model, better neural network structure built optimization. study results show that model has smaller root mean square error (RMSE) than human empirical which allows rapid accurate inversion parameters even noisy interference.