Theory and Applications for Wave Scattering in Tunable Complex Media - Philipp del Hougne (IETR, Rennes)

In this talk, I will review our recent efforts (i) to formulate physical models for wave scattering in tunable complex media, (ii) to frugally estimate the model parameters in unknown experimental settings, and (iii) to derive wave control protocols for previously unimagined or not-fully-understood scenarios. All results are corroborated with microwave experiments based on a chaotic cavity parametrized by a programmable metasurface, in line with the emerging “smart radio environment” paradigm, but the concepts’ generality extends to other scales and wave phenomena. In the first part, I show that compact physical models based on a coupled-dipole formalism or multi-port network theory can be formulated without an explicit description of the overwhelmingly complex medium. In the second part, I show that the model parameters can be estimated via gradient descent or in closed form, usually with ambiguity which facilitates the parameter estimation. I also clarify under which conditions the ambiguities can be (almost) eliminated. Moreover, I highlight that the model parameters can be estimated purely based on phase-insensitive measurements thanks to the constraining mathematical structure of the physical models. In the third part, on the one hand, I demonstrate how to tune an unknown chaotic cavity to feature a coherent perfect absorption state at a desired frequency and how to identify the wavefront that excites that state purely based on non-coherent measurements, unlocking a previously unimagined regime of frugal wave control limited to non-coherent detection. On the other hand, I demonstrate how to achieve optimal non-invasive focusing on a moving target inside an unknown complex medium, highlighting in particular that two target displacements are required (which had not been understood in related prior studies). I also extend this concept to optimal non-invasive time-averaged power delivery to a perturbing target inside dynamic complex media.

Contact: Dorian Bouchet