Analysis of lossy systems via 2D-FDTD and ESPRIT

The two-dimensional finite difference time domain method has been introduced as an efficient full-wave method for waveguide structures which are uniform in the direction of wave propagation. Liu, Schutt-Aine and Chen (see IEEE AP-S, p.1074-1077, 1999) used signal processing techniques such as ESPRIT as a post-processing engine to extract the frequency domain information from the sampling data. Due to the consideration of the analytical nature of the solution, this technique reduces the computational time dramatically. A combination of the 2D-FDTD and signal processing techniques reduce the memory size and CPU time at the same time. We extend the algorithm of Liu et al. to lossy systems. The dielectric and conductor losses are both included in one analysis. The super-resolution algorithms from signal processing techniques are employed in the temporal domain to extract the parameters from the sampled data in a shorter simulation.