Statistical modeling of antennas via a generalized stochastic-FDTD method

The consistent analysis of complex antenna setups with statistically varying media attributes is performed in this paper through a 3-D curvilinear stochastic finite-difference time-domain technique. The new method generalizes the concept of covariant/contravariant metrics for curved structures and derives the mean value and standard deviation of field components in a single run. Also, for the acceleration of the resulting simulations, all codes are developed in terms of graphics processor units and parallel programming. The profits of the proposed algorithm over multiple-realization Monte-Carlo schemes are certified via comparisons from circular- and trapezoidal-toothed log-periodic antennas with various realistic uncertainties in their materials.