This paper proposes a wideband measurement technique for the simultaneous characterization of complex permittivity and complex permeability using an open-ended coaxial probe (OECP) for in situ applications. The method combines a wideband deep neural network (DNN) with the two-thickness method (TTM) to obtain two independent reflection coefficients, enabling the determination of both complex permittivity and permeability across a broad frequency range. The wideband DNN is trained using a dataset generated by finite element method (FEM) simulations, allowing the approach to handle challenging measurement scenarios, such as small-sized flanged OECPs, which are often difficult or impractical to analyze using the conventional spectral-domain method (SDM). Numerical and experimental results demonstrate that the proposed technique can achieve higher accuracy in broadband EM parameter extraction compared with SDM, while significantly reducing in situ computational time, making it highly suitable for nondestructive, real-time characterization of electromagnetic absorbers and related materials.
Chen et al. (Thu,) studied this question.