A MEASUREMENT-BASED MODEL FOR PREDICTING COVERAGE AREAS OF URBAN MICROCELLS

We have performed data reductions on 900 MHz signal attenuations measu red on numerous streets in Manhattan. The data base consists of both l ocal spatial averages of signal attenuation and the short-term fluctua tions about this average. The former, which we call the local mean att enuation (LMA), is our primary focus. We use the data base to obtain c ontours of constant LMA for two neighborhoods and find the contours to have the shapes of convex diamonds. We propose that squares inscribed within these contours be used as the ''building blocks'' of microcell ular environments. A theory is developed that explains the contours an d predicts, with reasonable accuracy, the sizes of the inscribed squar es. We show, moreover, that the prediction method can be applied witho ut the need for measured data. Finally, we examine the short-term fluc tuation statistics of the signal attenuation, showing them to be Rayle igh-like in the non-line-of-sight regions of a microcell and Rice-like in the line-of-sight region. We discuss possible extensions of this w ork to other frequency bands and other urban environments.