The statistical distribution of amplitudes and phases of paths are investigated for Time-Variant Impulse Responses that have been measured in 63 different scenarios in indoor environment. Each measurement consists of 89 consecutive impulse response samples. In contrast to investigations by other authors the occurrence of time-variant changes of path lengths has been taken into account for the detection of paths by applying an algorithm denoted as "path-tracking". Evaluation is done by the examination of frequency distributions for the remaining errors when fitting theoretical curves to distributions from measurements. A Weibull distribution shows the best fit for the distribution of path amplitudes. The results for the Nakagami distribution and the Rice distribution turn out to be rather similar to the results for the Weibull distribution, whereas Rayleigh and lognormal distribution seem to be less appropriate. The path phases are shown to be extremely good uniform distributed. A modeling approach that takes into account deterministic influences, especially the correlation between adjacent impulse responses, is also briefly discussed.

No abstract

No abstract