Statistical analysis of the influence of the morphology onto the whole-body exposure to radio-frequency waves

The increasing use of wireless technologies and the multiplicity of the usages arouse a lot of questions from the general public about possible health effects of these technologies. The influence of the human variability, in particular the morphology, onto the exposure induced by electromagnetic fields is one of the important topics. The development of high performance computing allows one to use nowadays the numerical dosimetry to assess the specific absorption rate (SAR) on which the associated uncertainty is a key question. But despite the large efforts that have been carried out over the last 20 years to develop computational methods as the FDTD, the evaluation of the SAR remains a time consuming procedure. Consequently, the uncertainty quantification in dosimetry is a true challenge and the propagation of the variability of the input morphological parameters through the physical model of the FDTD cannot be addressed with classical methods such as the Monte Carlo simulations. Advanced surrogate modeling techniques such as the polynomial chaos expansion (G. Blatman and B. Sudret, Journal of Computational Physics, 230(6), 2011, pp. 2345–2367) have to be used.