Measurement-based Path Loss Prediction and Analysis in Water Scenarios

Unmanned aerial vehicles (UAVs) have enabled the unmanned management of water areas. Path loss (PL) serves as a critical factor that affects the UAV communication quality. In this work, a PL model for UAV channel in water scenarios is proposed, aiming to explore the multidimensional relationship between transceiver distance, carrier frequency, UAV height, and PL by constructing an Artificial Neural Network (ANN). Moreover, channel measurements are conducted in a river area, and the obtained data are analyzed using the proposed ANN model as well as the close-in (CI) model. The analysis results demonstrates that the PL increases initially and then declines as the UAV height increases. The proposed ANN model has higher prediction accuracy than the classic empirical model, and can achieve precise prediction of the PL in water scenarios.