FDTD Modeling of Indoor Propagation with Frequency Averaging

The presence of numerical dispersion and anisotropy in the finite-difference time-domain (FDTD) method leads to an error in the phase of the numerical waves which in turn degrades the accuracy of the simulated electric field (E-field). In this paper, the error in the mean magnitude of the numerical E-field obtained by averaging over frequency is investigated. The central frequency, at which the study is performed, is 3 GHz and different averaging bandwidths are tested to check the accuracy of representing the large-scale fading. By comparing with measurements, it is found that before averaging the maximum error in the numerical results is of 64% for empty room and 57% for office scenario. However, the maximum error in the mean magnitude of the FDTD E-field, when averaging over bandwidth of 200 MHz is employed, is of 36% and 45% for empty room and office scenario, respectively.