Real-time prediction of W-band dynamic rain attenuation based on chaotic mapping particle swarm optimization BP neural network-ARIMA model

Abstract

Aiming at the low prediction accuracy, stability and real-time performance of ARIMA (1,1,5) rain attenuation model, which is commonly used to verify W-band Rain Attenuation Prediction model, this paper firstly improves the BP neural network model by using chaotic mapping and particle swarm optimization algorithm, and then combines the improved BP neural network with ARIMA (1,1,5) model to propose chaotic mapping particles. Group optimization BP neural network-autoregressive summation moving average rainfall decline prediction model. Then, the simulation results are compared with ARIMA (1,1,5) and ITU-R models respectively. The results show that the accuracy and stability of the model are better than ARIMA model, which shows that the model can be used for real-time prediction of W-band rain attenuation.