Research On Rain Attenuation Mitigation Models For Airborne Satellite Communication Systems And Methods For Satellite Communication Resource Allocation

With the rapid development of satellite communications,airborne satellite communication systems have been widely used due to their characteristics of global coverage and instant communication.On the one hand,as the communication frequency band increases,the rain fade of airborne satellite communication links also increases.Traditional rain fade prediction models have defects such as unified parameters,limited application range,and insufficient accuracy,making the study of more accurate rain fade models adaptable to different scenarios critically important.On the other hand,with the development of high-throughput,multi-beam satellites,traditional fade resistance technologies have become economically impractical.Moreover,the number of connected devices and the high traffic demands of end-users are increasing day by day.There is an urgent need for a rain fade compensation method that can meet user traffic demands while maximizing the utilization of system resources.Addressing the issue of rain attenuation prediction models,our research introduces a GRU based model.According to the ITU-R P.1853 recommendation,we synthesized rain attenuation time series data from six ground stations in China.We used these data sets to train and validate both single-layer and multi-layer GRU models.Our simulation results indicate that the multi-layer GRU model consistently outperforms the singlelayer model across three key metrics: root mean square error,mean absolute error,and the coefficient of determination.On average,the multi-layer GRU model achieves an accuracy improvement of 5.54% and 9.16% over the single-layer model in the training and validation phases,respectively.As a result,we selected the multi-layer GRU model as the final model for our study.The prediction accuracy of this model is approximately90.21%,demonstrating a relative improvement of about 2.92%,5.64%,and 6.53%compared to LSTM,SML-GPR,and XGBoost models,respectively.In response to the limitations of conventional power margin rain attenuation compensation methods,we conducted a study on resource allocation methods and proposed a resource allocation algorithm grounded in multi-objective optimization.This algorithm is designed to maximize system resource utilization while meeting user demands.It also alleviates interference between beams through the application of precoding techniques.During experimental evaluation,we set up four simulation scenarios: full precoding,partial precoding,no precoding,and the effect of rain attenuation.Simulation results reveal that while the no precoding scenario has a shorter calculation time,it fails to meet user demands as traffic intensity increases.The full precoding scenario offers high system resource utilization but is computationally extensive.The partial precoding scenario presents a balanced performance between computation time and system resource utilization across various traffic demand distribution scenarios.Under rain attenuation conditions,the system requires allocating more resources to rain-affected areas.As a result,compared to clear sky conditions,system resource utilization increases,and some user demands may remain unfulfilled.