Channel extrapolation based on wideband MIMO measurements

Abstract

This paper shows results for channel extrapolation in both spatial and frequency domain, based on real measurements in an indoor office environment. Using a network analyzer, and measurement robots with very precise movement steps, channel data over a frequency span of 50MHz has been colleted in a large spatial volume. By using some key parameters that are extracted from a subset of the data, together with a model of the channel, an extrapolated estimate of the channel at a second location is created. This estimate is then compared to the measurement that is available for the second location, which thus gives us a measure of how well channel information may be extrapolated in space and frequency. We show that by using a well assumed model of the propagation environment, we may create an extrapolated estimate of the channel for a different location, that can be used to increase various system performance measures like received power, well beyond the channel decorrelation distance.