A combination of FD-TD and Prony's methods for analyzing microwave integrated circuits

Efficient modelling and sensitivity analysis of lossy structures using FDTD

“…The auxiliary function given in (5) can be exploited to extract the complex propagation constants. Curve fitting approaches such as Prony's method can be utilized [6].However, the matrix pencil method (MPM) has recently show good accuracy and high efficiency for solving this problem [7]. Thus we exploited the adopted version of the matrix pencil method described in [8] to obtain the complex propagation constant i m γ for each mode and the amount of power coupled to this mode through the coefficients i m c .For this purpose, a 3D simulation of the cross section of the waveguide is performed.…”

Low-order models from FD-TD time samples

“…One drawback of the method is a long computation time while modeling structures with high quality factors. In order to circumvent the problem signal processing techniques, such as all-pole auto-regressive (AR) [2], zero-pole auto-regressive moving-average (ARMA) [1], [2], and various variants of Prony's [3]- [6] methods, have been proposed to extract time signature features from a short segment of the original FD-TD sequence. Excellent results have been reported for most of these techniques, but each of them involves a nonobvious and subjective selection of key parameters.…”

Adaptive Application of Autoregressive Moving Average Method in Accelerating FDTD Simulation

“…To reduce the computation time of FDTD simulation, method that can predict the future response, or that can extract the system transfer function from a short record of past time signals, should be applied. To meet this end, many signal processing techniques have been applied to extrapolate the FDTD signals, such as Prony's method [1], the autoregressive (AR) method [2], the generalized pencil of function (GPOF) [3], and the autoregressive moving average (ARMA) method [4].…”