Feasibility Analysis of the Prediction of High Power Electromagnetic Environment from Antenna Current Response

Abstract

How to predict the characteristics of high power electromagnetic (HPEM) environment surrounding the electronic victims is of vital importance for taking effective activities to be protected. Hence, this article tried to utilize the measured antenna response data excited by HPEM to predict the electric field wave of HPEM according to the combination of the system identification model and artificial intelligence method. The train and test data are generated by MoM method, which takes simple line antenna as the targeted antenna and double-exponential waves with multiple parameters as HPEM environments respectively. The classical system identi-fication method output-error (OE) model and nonlinear autoregressive model with external input (NARX) neural network (NARX-NN) are introduced to establish the inverse model of the trans-fer function from current response to HPEM. The OE prediction accuracy 68% and NARX-NN's 72.4% demonstrate that it is feasible to predict HPEM environment by using antenna response test data.