基于BP网络的卫星通信雨衰预测模型

电磁波的降雨衰减对卫星通信链路具有很大的影响。在分析基于BP算法的多层前馈网络具有逼近任意连续非线性函数的能力和降雨率与降雨衰减的非线性关系的基础上,提出并建立了基于BP神经网络的卫星通信降雨衰减预测模型,根据某地区的实测数据进行训练和预测,并与ITU-R模型进行了对比。结果显示神经网络模型具有较小的误差,为降雨衰减预测提供了一种更加准确有效的方法。

Rain Attenuation Prediction Model of Satellite Communication Based on BP Neural Network

Rain attenuation affects the performance of satellite communication link greatly. Considering the characteristic of approaching any consecutive nonlinear function of MuLti-layer Feed Forward Network based on Error Back-Propagation and the nonlinear influence of rain rate on rain attenuation, we propose a new rain attenuation prediction model based on BP neural network, perform training and prediction according to the actually measured data of a certain area, and make comparison with ITU - R model.The results show that the proposed model has little errors. The work in this paper proposes a more accurate and efficient method for rain attenuation prediction.