Multiple-Input Multiple-Output Channel Models: Theory and Practice

6

EXPERIMENTAL VERIFICATIONS

In this chapter, we present an experimental analysis of two correlative channel models; namely, the structured and Kronecker models. There is only a handful of correlative wideband MIMO channel models in the literature. As previously pointed out, the wideband Kronecker model is by far the most popular, mainly due to the separability assumption and therefore the simplicity of the model. We compare the performance of the structured versus Kronecker models using real-life channel data, gathered from the wideband MIMO software defined radio (WMSDR) and Brigham Young University (BYU) wideband channel sounder.

The chapter begins with a discussion of a few metrics commonly used to validate the performance of MIMO channel models. It then describes the experimental setup for the WMSDR and BYU data sets. Both sets of data were collected in very different environments, using two apparatuses whose specifications differ greatly. The WMSDR data were recorded in a fixed point-to-point outdoor environment. For most locations, the receiver had direct line-of-sight to the transmitter. In contrast, the BYU data were recorded using a much higher bandwidth than that of the WMSDR. The BYU data were recorded in a fixed point-to-point indoor environment, where no line-of-sight existed between the transmitter and receiver. We show the robustness of the structured model using both of these data sets.

Using the MIMO azimuth power spectrum (APS), we qualitatively show how the ...