Predicting the Electromagnetic Characteristics of an Unknown Antenna Using Noisy or Phaseless Sparse Data

Identifying the electromagnetic characteristics and vulnerability of unknown antennas is crucial for efficient electronic warfare, electromagnetic attacks, and antenna system defense. To accurately predict the overall electromagnetic characteristics of an unknown antenna, it is imperative to gather and utilize various indicators obtained by aerial reconnaissance. Based on the acquired magnitude and phase data of the sparsely radiated electric field, the antenna can be modeled using a linear fitting algorithm on a uniform three-dimensional grid. However, in practical measurements of radiation patterns by aerial reconnaissance, significant noise generally contaminates the observed magnitude and phase data to render it ineffective. This results in typical scenarios wherein only magnitude data are observed owing to various environmental variables or system-related factors. Therefore, further research is required to determine the approach to performing infinitesimal dipole modeling (IDM) using incomplete data. This study analyzed the performance of IDM in the presence of noise-contaminated magnitude and phase. It proposes a method for performing IDM using a non-linear optimization algorithm only for magnitude information. The algorithm is based on the phaseless data of collected sparse radiation patterns. We employed a levenberg-marquardt algorithm for non-linear optimization. For both simulations and measurements, a rigid horn antenna with a center frequency of 2.4 GHz was utilized. It was observed that increasing the noise ratio in sparse data caused an increase in errors. Even in the worst-case scenario of phaseless data, the nonlinear optimization-based IDM demonstrated remarkable recognition performance.