A New Nonstationary Gaussian Noise Model for Indoor Wireless Channels

In this paper we examine a non-stationary statistical model of the broadband response of an indoor wireless channel. The channel impulse response is comprised of a line-of-sight component and multipath components modeled by Gaussian noise weighted by exponentially decaying function of time. The aforementioned model is applied to evaluate the variability in the energy across the spectrum at the receiver. The model exhibits temporal non-stationarity in its variance and favorably captures the variability of the spectral amplitude observed in numerical simulations. It is shown that the standard deviation of the logarithm of normalized power spectral amplitude tends to 5.54 dB as signal to noise ratio approaches zero