Effects of User Body Shadowing in Indoor Wireless Channel Based on Ray Tracing

Abstract

In this dissertation, the effects of user body on radio wave propagation in indoor wireless channels are analyzed. The user who is nearly always being close to mobile device influences very strongly and consistently on propagation channel. Therefore, exclusively focusing on the user body separately from other bodies, the shadowing effects caused by the user are investigated at 2.4 GHz by using the uniform theory of diffraction (UTD) and the ray-tracing technique. First of all, the user body shadowing (UBS) effects on a single ray path are investigated deterministically by using the UTD. The UTD scattering solutions for diffraction at a smooth convex surface are adopted to analyze the effects of user body modeled as a circular cylinder. The UTD-based model for a single ray path is defined as the relative received signal power according to the relative position of user, which is validated by measurements in an anechoic chamber. The validated UTD-based model is combined with the indoor ray-tracing technique in order to examine the UBS effects on multipath channels. Since the ray-tracing provides not only the powers of multipaths but also their angular profiles, it is possible to apply the UTD single path model according to the relationship between the users position and the direction of rays. This combination method is also verified by in-building measurements. In realistic communications, however, the users position can be neither fixed at any one value and nor can its exact value be provided to systems in real time. Thus, a statistical analysis for the UBS is conducted taking into consideration the randomness of users position. First, the K-factor, defined as the ratio of the power in the dominant path and the sum of the powers in the other paths, is proposed as the most significant factor to determine the UBS effects. Because the UBS effects considerably depend on the extent of the dominant path and whether the dominant path is blocked. As a result, the distributions of total power losses caused by the UBS are link-by-link modeled by Nakagami-m distributions. Additionally, the estimated parameter m is proposed as a function of K-factor. Finally, the enhanced UBS stochastic model is proposed based on the bimodal characteristics of UBS. The UBS model based on Nakagami-m distribution has a drawback of inaccuracies for the links with high K-factor because the distribution of total UBS losses for links with high K-factors has a bimodal shape that has two peaks in its histogram. Therefore, the distributions of total UBS losses were classified into unimodal and bimodal groups with the quantitative decision criterion of K-factor. For the unimodal model, Rician distribution is used to achieve the best accuracy, whereas Gaussian mixture model is exploited for the bimodal UBS model. The validity of these proposed models is verified using the ray-tracing simulation in various indoor environments.