Principal Component Analysis of Results Obtained from Finite-Difference Time-Domain Algorithms

Finite-Differences Time-Domain (FDTD) algorithms are well established tools of computational electromagnetism. Because of their practical implementation as computer codes, they are affected by many numerical artefact and noise. In order to obtain better results we propose using Principal Component Analysis (PCA) based on multivariate statistical techniques. The PCA has been successfully used for the analysis of noise and spatial temporal structure in a sequence of images. It allows a straightforward discrimination between the numerical noise and the actual electromagnetic variables, and the quantitative estimation of their respective contributions. Besides, The GDTD results can be filtered to clean the effect of the noise. In this contribution we will show how the method can be applied to several FDTD simulations: the propagation of a pulse in vacuum, the analysis of two-dimensional photonic crystals. In this last case, PCA has revealed hidden electromagnetic structures related to actual modes of the photonic crystal.

Signal Processing Techniques for Chipless UWB RFID Thermal Threshold Detector Detection.

This letter presents signal processing techniques to detect a passive thermal threshold detector based on a chipless time-domain ultrawideband (UWB) radio frequency identification (RFID) tag. The tag is composed by a UWB antenna connected to a transmission line, in turn loaded with a biomorphic thermal switch. The working principle consists of detecting the impedance change of the thermal switch. This change occurs when the temperature exceeds a threshold. A UWB radar is used as the reader. The difference between the actual time sample and a reference signal obtained from the averaging of previous samples is used to determine the switch transition and to mitigate the interferences derived from clutter reflections. A gain compensation function is applied to equalize the attenuation due to propagation loss. An improved method based on the continuous wavelet transform with Morlet wavelet is used to overcome detection problems associated to a low signal-to-noise ratio at the receiver. The average delay profile is used to detect the tag delay. Experimental measurements up to 5 m are obtained.