A Method for Evaluating the Directionality of the HPM Helical Array Antenna at Varying Input Power Levels with Multilayer Perceptron

In the target positioning of high-power microwave (HPM) systems, the measurement of the pointing angle of the high-power microwave antenna plays a key role. In this paper, we propose a novel multilayer perceptron (MLP)-based method for determining the pointing angle of the high-power microwave helical antenna. This method can predict the pointing angle even when there are errors in the input. To validate our idea, we constructed a dataset that includes the real and imaginary parts of the voltage signals received by the dipole antenna, along with the corresponding output angles. We used this dataset to train our MLP and demonstrated its effectiveness, showing that the MLP can produce accurate results even for input data not present in the training set, in contrast to traditional methods. Moreover, the MLP exhibits enhanced representation and generalization capabilities.