Microwave characterization using partial least square regression

Inverse problems for determination of dielectric materials properties (complex permittivity) are usually solved by iterative methods using numerically based forward model. These methods are computationally expensive. In this paper, we propose a fast inversion model based on partial least square regression (PLSR). The idea is to build a model able to predict in real time the properties of the sample under test using measurements of admittance or reflexion coefficient of the propagating electromagnetic micro wave along the coaxial line. Numerical solution of the direct problem is made using Finite Element Method (FEM). © 2016 IEEE.