Generalized Gamma-Laguerre Polynomial Chaos to Model Random Bending of Wearable Antennas

Abstract

A novel generalized Gamma-Laguerre polynomial chaos expansion is proposed to account for the effect of random variations in lower bounded design parameters on antenna performance. After fitting a shifted generalized Gamma distribution to datasets of such random variables, a predistorted polynomial chaos expansion is generated based on a set of orthogonal generalized Laguerre polynomials. The new statistical methodology is applied to assess the random change in resonance frequency when bending a wearable antenna around different parts of the human body, such as a leg, an arm, and a head. For different datasets, an excellent statistical fit is found to both the estimated probability density function of the bending radius and the resulting statistical distribution of the resonance frequency, while requiring up to 480 times fewer sample evaluations.