Impulse response modeling of indoor radio propagation channels

If indoor radio propagation channels are modeled as linear filters, they can be characterized by reporting the parameters of their equivalent impulse response functions. The measurement and modeling of estimates for such functions in two different office buildings are reported. The resulting data base consists of 12000 impulse response estimates of the channel that were obtained by inverse Fourier transforming of the channel's transfer functions. It is shown that the number of multipath components in each impulse response estimate is a normally-distributed random variable with a mean value that increases with increasing antenna separations; a modified Poisson distribution shows a good fit to the arrival time of the multipath components; amplitudes are lognormally distributed over both local and global areas, with a log-mean value that decreases almost linearly with increasing excess delay; for small displacements of the receiving antenna, the amplitude of the multipath components are correlated; the amplitudes of adjacent multipath components of the same impulse response function show negligible correlations; and the RMS delay spread over large areas is normally distributed with mean values that increase with increasing antenna separation.<
>