Institut Fresnel

We are building and exploiting several microwave measurement tools. In particular,
we have access to two anechoic chambers where we measure the electromagnetic field radiated by antennas in a controlled environment and/or the field scattered by complex targets. The measurements can be made in a monostatic configuration or a bistatic configuration
for a frequency range going from 700 MHz to 26 GHz.
The large chamber provides a spherical measurement geometry and has been used to measured the scattered fields of the Fresnel database published under three special sessions in Inverse Problems journal. In the small anechoic chamber, which has dimensions of 3mx3mx3m, the antennas can move in an horizontal plane, which is particularly convenient to probe buried targets inside a tank filled with an inhomogeneous material.
We also exploit a circular microwave scanner, which works at a given frequency of 434 MHz and where we can measure in a bistatic configuration the field scattered by targets placed within a circular tank. The overall diameter of this setup is of 60 cm.
We are also building dielectric constants characterization tools for samples of either solid or granular materials.

We are developing simulation tools in order to model the interaction of the electromagnetic fields with their environment. In particular, we are working on finite element codes either in 2D or 3D.

Optimal design is a key-tool for us in order to provide the best microwave setup for the envisaged applications. In particular, we want to determine the minimal number of antennas, their
adequate positions, the best frequency range, ... in order to get measurement which are robust to noise and non-redundant. To this hand, we are developing theoretical tools in order to answer to these questions, based on the spectral properties of the scattering operator. This research area help us to define robust and simple calibration procedures.

By probing the environment with electromagnetic waves, we are interested in retrieving
the characteristics of the illuminated targets. These samples can be manufactured and placed inside a complex environment (pipes, anti-personal mines). These samples can also be of heterogeneous nature (soil moisture, body temperature, ...).
To retrieve the permittivity maps of such targets, we are developing inversion algorithms, such as qualitative reconstructions algorithms (backpropagation, tomography, DORT, Field correlation techniques, ...) or quantitative iterative reconstruction algorithms (Newton methods with adjoint fields). A priori information can also be introduced in order to enhance the associated results (level-set algorithms, basis functions representation, regularization, ...)