A Novel LoS Probability Prediction Model for UAV-to-Ship Communications in Maritime Scenarios

The optimization of line-of-sight (LoS) probability is essential for enhancing 6G unmanned aerial vehicle (UAV)-to-ship (U2S) communication in maritime scenarios, where the communication channel is affected by dynamic sea waves (SW) and six-dimensional (6D) transceiver positioning. Traditional models have been found inadequate in addressing these variables. In this paper, a novel LoS probability model is introduced, integrating dynamic SW and 6D positioning of ships and UAVs. The effects of SW height, number, and period on LoS blockage are quantified, providing a tailored approach for maritime environments. A dual-channel convolutional neural network (DCCN) coupled with spider monkey optimization (SMO) is applied to refine the complex relationship between environmental factors and LoS probability. The model, trained with ray tracing (RT) data, is shown to significantly outperform standard models like 3GPP and 5GCM, offering new insights into U2S communication in maritime scenarios.