Abstract:
The performance of modern wireless systems within buildings is primarily limited by interference from other users and other systems operating in close physical proximity. In order to estimate system performance in the presence of interference, reliable models to predict signal strengths (and other channel parameters) are required. Currently, empirical models based on experimental measurements are widely used. However, these models inherently lack an electromagnetic or physical basis and can vary considerably between buildings. Indoor environments can be very complex, and improved accuracy and a more thorough understanding of how the radio waves propagate may be gained through an electromagnetic approach. However, the high computational costs of fully-electromagnetic techniques, such as the Finite-Difference Time- Domain (FDTD) method, limits their applicability for day-to-day system planning purposes at this stage. The research presented in this thesis describes a series of FDTD investigations of the indoor radio channel, using two- and three-dimensional geometries, with the aim of identifying the dominant mechanisms governing propagation within buildings. Mechanistic models are proposed by developing computationally efficient approximations for each mechanism. The complexity of the propagation processes can make conclusive identification of the mechanisms difficult, however various visualisation techniques developed in this thesis can be used to infer the dominant propagation paths. Central to the development of mechanistic models is a through assessment of the limitations of the FDTD method to model propagation within buildings, particularly the use of two-dimensional representations of the geometry, the impact of furniture and other details, and comparisons with experimental measurements.
