A Prediction Method Of Field Intensity Distribution Based On Finite Integral Method And Machine Learning

In this paper,the most common methods of wave propagation prediction currently are analyzed with their advantages and disadvantages compared.Similarity exists in the process of all rays intersecting with the triangle surfaces,which can be mined with machine learning.In this paper,a hybrid method of field strength prediction combining finite integral method and machine learning is proposed.To explore the similarity,basic unit of terrains and buildings(triangle surface)is built in CST.Then the electromagnetic field distribution of the incident triangle surface is calculated by simulation.To find out the factors influencing the field distribution,we repeat the simulation under different input condition and get the outputs respectively.The space around the triangle surface is divided into the near field and the far field.In the near field,we predict the field strength through the machine learning model.And for the far field region,the process is divided into two steps:first,predict the field information on the interface through the machine learning model,and then calculated the Poynting vector of the corresponding point in order to get the position of the wave propagation direction.At last,predict the field strength by using ray-tracing method in the far field.This paper makes simulations for the new method,we regard the plane wave irradiating on a triangle face as an example,compared prediction results of four kinds of machine learning model in near field,and then compared the simulation results of CST and hybrid model in far field,the results show that the hybrid model can ensure the accuracy of the conditions,and at the same time,this method can improve the computational efficiency.