Page 260 - Integrated Wireless Propagation Models

238 C h a p t e r F o u r

and tighter grouping around zero are shown for the two optimized models. The histo­
gram shapes of two predictions are the same, but the difference of about 20 dB is the
loss due to the building blocks, as mentioned before. When using the Lee microcell
prediction model instead of the Lee default model to compare with the tuned Lee model,
the results shown in the two figures will be almost identical.
The performance demonstrated in Figs. 4.4.3.1 and 4.4.3.2 clearly demonstrates the
importance of measurement integration and the Lee microcell model as a suitable can­
didate for propagation prediction in dense urban environments. When the building
data are available, the building block loss L8 will be obtained from the curve shown in
Fig. 4.2.1.1.2. When the building data are not available, simply set the block loss L8 "' 20 dB
in the Lee microcell model to get a fairly accurate prediction.

4.5 Other M i croce l Prediction Models
l

4. . 1 I n troduction
5
In this section, several approaches of modeling propagation in a microcell environment
are introduced. Both theoretical and empirical methods have been used, and ray-tracing
techniques have also been investigated. The multipath effects in the prediction results
are very important in urban areas, depending on the relative height of the base station
antenna and the surrounding buildings.
Generally, all existing microcellular models are valid only in flat urban areas, and
little attention has been given to the influence of terrain variations; the effects of vegeta­
tion have also been largely ignored. Both of these aspects are important and need to be
incorporated specifically into the ray-tracing models.
For the practical application of microcell propagation models, an important trade­
off is between the accuracy of the prediction and the computational speed with which
the prediction can be obtained. Microcells often have to be deployed in the field very
quickly, with little engineering effort. The guiding procedures and rapid statistical
planning tools are very important. A very high resolution topographic database is
required.
These microcell models are used and run at the start of a system deployment, and
then used to create a unique set of predictions and recommendations for deployment.
The real-time processes are operating within the base station, with assistance from the
mobiles, which can optimize and are used by the system to assess the likely system
parameters, such as transmit powers, antenna patterns, and channel assignments, on an
ongoing basis. This section describes models other than Lee model that can also yield
reasonable prediction accuracy when their parameters are tuned against measurements.
4.5.2 Empirical P ath Loss) Models
(
Normally, microcell models are based on statistical techniques. The Lee model is one
kind of statistical model.
Empirical models are statistical models with lower computational overhead but
less accuracy in small urban cells and subjective clutter classification. The parameters of
the model are the properties of the buildings along the radio path between the transmit­
ter and the receiver. These models are not as accurate for smaller cells under 1 km.
However, these models should not require any calibration and are more suited for the
urban macro- and microcells.