Prediction of Indoor Millimeter Wave Propagation Characteristics Based on BP Neural Network

Abstract

Due to the complexity of the millimetre-wave indoor propagation environment, using the conventional methods to analyze propagation characteristics of radio wave leads to great error, and the computational complexity is high. In view of the above problems, the prediction method for propagation characteristics of the indoor millimetre wave based on BP neural network is proposed; using its function approximation to describe the millimetre wave propagation parameter, finding out the mathematical relationship between its corresponding external factors and the signal attenuation, the mathematical logic relationship is constructed through establishing the prediction model. Simulation results show that using this method can quickly and intuitively obtain the relationship between the propagation distance and field intensity or path loss, respectively, with advantages of taking short time and high computational efficiency.