Modelamento de sistemas móveis celulares em ambiente indoor usando equações parabólicas

Abstract

The progress of the wireless digital communication system during the last years has been remarkable, and today is one of the fields with faster development. There is a growing search in propagation models that provide efficiency and accuracy. The study of the radio propagation in these environments is very important for the determination of the propagation losses in indoor environment, because any object with dimensions in the order of greatness inside of the range of mobile communication wavelength can act as an interference source for the electromagnetic waves. Two types of propagation models are very used for indoor environment: the empirics and the deterministic. The first ones are based on very simple and direct formulas, easy to apply, however they provide local-specific precision. The deterministic models follow propagation of electromagnetic waves physical principles - the most popular are the ray tracing and the numeric solution of the equations of Maxwell (elliptic equations and for approach, parabolic equations). In this work is proposed a model based on the method of parabolic equations, being considered narrow angles (up to 15°) and wide angles (up to 90°) of propagation. For the solution of the parabolic equation for small propagation angles, the implicit finite difference scheme of the Crank-Nicholson type was used; and for the solution of the equation that considers wide propagation angles the mixed Fourier Transforms was used. In order to validate the model proposed, measurement campaigns were carried out in frequencies of 900 MHz and 2.4 GHz and the data obtained from these campaigns were compared to the model proposed in this work. Besides, the model in subject was also compared with some existent models in the literature. The results found in this work validate the method of parabolic equations as an important tool for the calculation of the propagation loss in indoor environment, once a great reduction of computational effort and quantity of necessary memory, when compared the other methods and a better result was observed for wide propagation angles.