Electromagnetic Models for Indoor Wave Propagation Analysis and their Application for Ultra-wideband Near-field Radar Imaging of Building Interiors and Human Movement Detection.

Abstract

An efficient and accurate hybrid method for the analysis of wave propagation in complex building environments based on full-wave solutions and physical optics is introduced. Contrary to standard wave propagation methods such as "ray-tracing", which can only handle homogeneous walls of large extent, this method can account for the special propagation effects of inhomogeneous periodic walls (like cinderblock or reinforced concrete) with finite extent and irregular shapes at the ultra-high frequency (UHF) band. The method is described and validated in two dimensions and extended to 3D. Through analyzing a realistic indoor scenario, the effect of inhomogeneous periodic walls on the wireless channel parameters and possible errors of a homogeneous wall assumption are pointed out.

This method is combined with a full-wave model of a human body to expand the method for analyzing ultra-wideband radar detection and tracking of humans within a realistic building environment. It is shown that the EM backscattering of a human inside building is so weak compared to the wall return that it can be indistinguishable from the building return. By taking advantage of the human movement, detection of a human inside buildings is shown possible. Furthermore, the analysis is extended to multi-static radar imaging and it is concluded from the simulations that is possible to localize a moving human with radar measurements form outside up to a certain depth inside a building, subject to possible additional false detections caused by the building multipath propagation. These simulation results are validated with actual radar measurements of a mannequin in a lab environment. At the end, spectral and wall estimation techniques are introduced to the radar imaging of buildings and moving humans inside to improve image quality and consequently to enhance possible detection and localization.