No abstract

The field strength level of received signal and its statistical distributions are important in the study of signal propagation and scattering for wireless communication and remote sensing. In this paper, the scattering by random rough surface is analyzed by solving the solution of Maxwell equation. The method of moments is used to discretize the integral equation into the matrix equation. The sparse-matrix canonical grid method, which is a fast matrix solver, is applied in the analysis. Conjugate gradient (CG) method is adopted to solve the solution of matrix equation. With the solution of Maxwell equations, the magnitude of the scattered field is then used to predict the field statistics. Both time harmonic scattering and ultrawide-band (UWB) scattering are considered. For the time domain response, the electromagnetic scattered field in the vicinity of center of rough surfaces is first calculated in the frequency domain. Then the time domain signal is obtained by mean of the Fourier transform. The fading statistics are compared with that of the Rayleigh and Nakagami distributions. Results reveal that UWB signal exhibits less fading than the narrow band signal.

We studied wave propagation in urban environment based on Monte Carlo simulations of Maxwell's equations. The multilevel steepest descent fast multipole method (ML/SD/FMM) is used in the computations. The building profile is about 2 km in length and the frequency is 900 MHz. Numerical results are statistically analyzed to study the spatial variations including fast fading, shadow fading, and range dependence. The results are compared with the Rayleigh, Ricean, and lognormal distributions. Index Terms-Channel modeling, fading, fast multipole method.

No abstract

The rapid growth of technology in computing architecture and networking has resulted in fundamental changes in the way we perform computational analysis in science and engineering. Some problems that are intractable because they are computation‐intensive in memory and/or CPU time can now be solved in an effective manner with the help of parallel computing. The aim of this article is to introduce to readers some examples of parallel computing algorithms that were developed for the computational electromagnetics community but are also applicable to other fields in science and engineering. In this article, the construction components of parallel computing platforms are introduced. A Beowulf cluster, which is a system with distributed memory multiprocessor architecture, will be briefly discussed. Some of the most popular simulation techniques in computational electromagnetics such as the finite‐difference time‐domain (FDTD) method and the method of moments (MoM) are discussed. Several methods of solving a full matrix either directly or iteratively will also be discussed. Illustrative examples are given to demonstrate the use of these algorithms and their implementations on the parallel computing platform for solving large‐scale electromagnetic problems.