A measurement-based model for predicting coverage areas of urban microcells

The authors have performed data reductions on 900 MHz signal attenuations measured on numerous streets in Manhattan. The database consists of both local spatial averages of signal attenuation and the short-term fluctuations about this average. The former, which is termed the local mean attenuation (LMA), is the primary focus. The database is used to obtain contours of constant LMA for two neighborhoods. It is shown that the contours have the shapes of convex diamonds. The authors propose that squares inscribed within these contours be used as the building blocks of microcellular environments. A theory is developed that explains the contours and predicts, with reasonable accuracy, the sizes of the inscribed squares. It is also shown that the prediction method can be applied without the need for measured data. The short-term fluctuation statistics of the signal attenuation are examined. They are shown to be Rayleigh-like in the non-line-of-sight regions of a microcell and Rice-like in the line-of-sight region. Possible extensions to other frequency bands and other urban environments are discussed