Page 180 - Integrated Wireless Propagation Models

158 C h a p t e r T h r e e

regarding how to apply this method to real system applications. Finally, suggestions for
future enhancements will be discussed.

M
3 . 2 . 1 The Terrain Normalization e thod
Before discussing normalization of the terrain, it is necessary to know exactly how the
terrain contour affects the prediction. The path loss curve constitutes a 1-mile intercept
P0 and the slope of the path loss y, as shown in Eq. (3.2.1.1). The method used to find
P0 and y for each geographical area is described below.
From Eq. (3.2.1.1), it is shown how the gains and losses vary with terrain along the
mobile direction. The purpose at this stage is to factor out the terrain variation and to
show the effects of human-made structures on the prediction. To achieve this, we need
to either nullify the terrain effect or do terrain normalization.
Slope y and intercept P0 are determined due to different human-made environ­
ments. Gains or losses calculated based on the antenna height gain use terrain contour.
As mentioned in the previous section, the measured data are affected by both human­
made structures and terrain contours. If we can separate the terrain variation effects on
the received signal from the human-made effects, then we can obtain the right set of
values for the slope and the 1-mile intercept for the human-made effects.
Now let us explain what exactly the terrain normalization does on the measured
data points. Every measured data point is considered individually, and the following
steps are carried out:

Step 1: Collect drive test data
Step 2: Screen the valid measured data
Step 3: Run terrain normalization on the measured data
Step 4: Calculate the slope and intercept based on data from step 3

In order to obtain meaningful statistics, a large amount of data needs to be collected.
The measured data need to be carefully screened because bad data can lead to a wrong
conclusion. The data collected in tunnels, elevated highways, or below the noise floor
need to be discarded.
The terrain normalization method in step 3 is discussed in more detail here. For
every measured point, the following steps need to be performed.
First, extract a terrain radial from a cell site from the terrain base to match the mea­
sured point. Then calculate the prediction along the radial. Depending on whether or
the path is being clear or shadowed, calculate both the gain or the loss at that point and
the adjustment factor:

=
P, P0 - y l og (!...) + Geffh - L(v) a
+
ro ;/ t '
(3.2.1.1)
Gain Loss Adjustment
By nullifying the effects of gain or loss and adjusting the measured data M from a
different setting, the fully normalized points from the measured data are then obtained:

(3.2.1.2)
where Mn m is the normalized signal from the measurement.
or