An Efficient Method of Two-way PE Based on ANN for solving field value in obstacle environment

Abstract

This paper proposes a method to efficiently calculate the spatial field value by using an artificial neural network (ANN) to approximate the complex propagation results inside the obstacle under the framework of the two-way parabolic equation (2W-PE) method. The previous algorithm takes into account the multiple reflections of the radio wave inside the obstacle and this method can calculate these propagation processes under the calculation framework of the 2W-PE by fitting with multiple ANNs, which greatly accelerates the computational speed of spatial field value prediction in the obstacle environment. In addition, due to the complexity of the environment, it is difficult to determine the boundary condition coefficients of obstacles in the calculation, so we apply a non-end-to-end trained ANN. The simulation results show that this method can greatly increase the calculation efficiency.