Antenna effects on indoor wireless channels and a deterministic wide-band propagation model for in-building personal communication systems

TR Number



Journal Title

Journal ISSN

Volume Title


Virginia Tech


While the application of antenna diversity in a narrow band communication system is well understood, little research has been done on antenna effects in wide-band channels. Research has shown that circular polarization (CP) is more robust in combating multipath than linear polarization in line-of-sight channels. One objective of this thesis is to study the effects of antenna polarization and antenna pattern on multipath delay spread and path loss in indoor obstructed (OBS) wireless channels. A wide-band experiment was performed in a two-floored modern office building at 2.4SGHz in August 1991. Some preliminary results are as follows. Circular polarization cannot reduce delay spread in OBS channels and CP signals are more vulnerable to depolarization in OBS channels. Our results show that vertically polarized (VP) directional antennas at both the transmitter and the receiver can give better delay spread and path loss results than other antenna combinations. The performance of VP directional antennas are found to be sensitive to the alignment of the antennas, and the performance gain over omnidirectional antennas degrades as shadowing effects increase.

In the second half of the thesis, a deterministic wide-band propagation model that can predict channel impulse responses inside buildings is proposed and implemented. The three dimensional image-based propagation model includes effects of antenna pattern, antenna polarization, geometry of the building, and building materials. Comparisons between measured and predicted power delay profiles are given in Chapter Seven. Preliminary results show that the worst case path loss error is IOdB, and the standard deviation of path loss error is 4.6dB. For most cases, predicted rms delay spread values are 20ns within the measured values. Possible prediction errors are due to unmodeled furniture inside the offices and limitations of Geometrical Optics (GO) assumptions. The algorithm is shown to be more efficient than brute force ray tracing algorithm if the number of objects are on the order of a few hundred. Acceleration techniques for the algorithm are also discussed in the thesis.