I. Introduction
THE finite-difference time-domain (FDTD) method has been widely used in the field of computational electromagnetics due to its simple implementation and a capability to address complex targets [1]–[3]. From the viewpoint of the moment method, the conventional FDTD is based on a rectangular-pulse field expansion [4]. Poor FDTD numerical dispersion properties are inevitable due to the discontinuous basis functions. As a result, field components have to be oversampled to obtain numerical results with reasonable accuracy. This requirement becomes computationally expensive for electrically large targets.