Page 342 - Integrated Wireless Propagation Models

320 C h a p t e r F i v e

certain measurement data are needed to ensure the accuracy of the model. There is always
a cost associated with accuracy. The model needs to be able to balance how detailed the
input data should be, how many measurement data need to be collected, and how accurate
and efficient the model needs to be. Usually, site-specific measurements (for point-to-point
models) are needed. The data can be used to determine the details of the propagation
mechanisms and material parameters for a particular building. These "measurement-based
2
prediction" approaches are described in detail in reference4 and early sections of this
chapter. The Lee picocell model is able to be fine-tuned for accuracy and efficiency from the
measurement data. The measurement data should be collected across entire buildings.
There are many different parts of empirical models for indoor propagation
prediction; most of which are based on the same propagation path-loss principles.
The first approach is introduced from the Keenan model, which uses the free space
curve as the baseline and adds additional loss factors relating to the number of floors
n and walls nw intersected by the straight-line distance r from the intersection to the
1
receiver.
The second approach is from the Ericsson in-building path-loss model.43, 4 4 It is similar
to the multiple-breakpoints approach discussed in Chap, 4 on the Lee macro-to-micro
integration model. There is a unique slope from one breakpoint to another breakpoint in
the Ericsson model. This approach allows the model to be flexible so that field measure­
ment data are gathered more efficiently and move accurately, However, this approach
tends to lead to large errors in some areas due to the large variability in propagation
mechanisms involved in different building types and in different wave paths within a
single building. Although the accuracy and flexibility can be improved by using the multiple­
radials approach as adopted by the Lee microcell model and discussed in Chap. 4, the
calculation of the Ericsson model becomes complicated and slow for in-building scenarios,
The third approach is from the empirical log-distance path-loss prediction
models,IM5 which should be used to determine a high-level system design. Once the
design is complete, the field measurements or an actual deployment can provide a
final adjustment to a model's applicability in a given building, A good static design
must still include a sufficient margin for deploying a system based on the environ­
mental dynamics, Depending on the bandwidth and data rate, delay spread effects
may also need to be considered, As a common practice, the delay spread effects are
considered by the radio specification of the receiver and is handled at the equipment
requirement,
The Keenan model provides some flexibility and is adopting the measurement data
to tune the model with accuracy and speed. The Ericsson multiple-breakpoint model
provides an empirical curve of worst-case attenuation versus distance. The attenu­
ation variable in the log-distance model provides some flexibility to the model. The
description of log-distance model is not covered in this book because the Lee model
does not use it for comparison, The readers can find the details of log-distance model in
1
references, 4,45

l
(
5.4. 1 The Motley-Keenan Model E mpirica ) and a Comparison
with the Lee Model
The Motley-Keenan model considers all walls that are intersecting the direct ray
between transmitter and receiver. The user can adjust the attenuations for the walls.
The formula for the modified Keenan-Motley model has been shown in Eq. (5.2.8.1 . 1 ) .