Date of Award


Document Type

Open Access Thesis


Electrical Engineering


College of Engineering and Computing

First Advisor

David W. Matolak


Modeling of indoor radio channels has been a dynamic area of research in recent years because of the increasing demand for indoor wireless communications. In order to efficiently deploy such indoor systems, a good knowledge of the indoor wireless channel’s characteristics are required. One important characteristic is an accurate propagation path loss model. For developing such models various measurements have been carried out. In modeling propagation path loss in complex indoor environments, accurate and fast algorithms for estimating the local mean signal level are essential; these are also of use for power control and handoff decisions. These local mean power levels are typically characterized statistically. This thesis seeks to add to the body of knowledge regarding indoor channels by determining important parameters for estimating local mean power in line-of-sight (LOS) conditions.

In this thesis, a set of indoor channel measurements was taken, and from these measurements, statistical channel characteristics were derived. These characteristics, or parameters, include the spatial averaging window length (2L), interval between spatial samples (dmin), and the number of sampling points (Nmin) within the window. These parameters are used in estimating the local mean value of a radio signal. Measurements were conducted in a LOS condition in a corridor at a frequency of 5.725 GHz, at nine different link distances (from 4.57 m to 41.15 m). A window based weighed sample average power estimator was used to determine the statistical characteristics of local average power at each link distance. Computer simulations were also performed in order to verify the algorithm’s performance before application to the measured data set. A set of relationships among 2L, dmin and Nmin were also established. It was observed that as the link distance increases in our LOS corridor environment, the number of sampling points, Nmin and averaging window length 2L tend to decrease. The decrease with link distance is not perfectly monotonic due to inhomogeneities in the environment, yet the general trend is as expected due to the wave guiding effect of the corridor. Future work in this area would include additional measurements in different corridors, and in different LOS indoor settings, with the aim of determining general guidelines for the estimation parameter values in such environments.