Understanding propagation model types

This section describes the propagation model types that Mentum Planet supports.
Slope-based models, such as the Okumura-Hata model, take clutter into account au
tomatically when generating predictions. Deterministic models, such as the CRC-P
redict model, depend on the model of the environment and the specification of cl
utter property assignments. Table 1 rates how each of the three main propagatio
n models perform when used under certain conditions.
Table 1 Ratings for popular propagation models
Used...
CRC-Predict
Planet General Model
Universal Model
For macro-cell planning
Good
Good
Excellent
For mini-cell planning (urban)
Poor
Fair
Excellent
For micro-cell planning (urban)
Very poor
Poor
Excellent
Over large propagation distances
Excellent
Fair
Good
With no model tuning
Fair
Poor
Good
With cluster tuning
Fair
Fair
Excellent
On a per sector basis
Good
Excellent
Excellent
With merged predictions
Good
Fair
Good
Free Space model
You can use the Free Space propagation model where line of sight situations exis

t with no Fresnel zone obstructions. For example, this model is useful for high
frequency, short distance, and Local Multipoint Distribution Service (LMDS) appl
ications.
The Free Space model is used for path loss estimation where there is an unobstru
cted line of sight between the transmitter and the receiver and there are no obs
tructions within the first Fresnel zone. This is often the case for satellite an
d microwave communications. The Free Space model is based on the Friis Free Spac
e equation, which states that the received power drops off and is calculated as
the square of the distance between transmitter and receiver (i.e., 20 dB/decade)
.
Okumura-Hata model
You can use the Okumura-Hata model for urban or suburban areas if little is know
n about the terrain and clutter.
The Okumura-Hata algorithm is entirely empirical. It is based on a multitude of
measurements from selected urban centers in Japan. Okumura developed a set of cu
rves giving the median attenuation relative to free space for an urban area of q
uasi-smooth terrain. Base station effective height varied from 30 meters to over
800 meters, and mobile antenna height was 3 meters and 1.3 meters, both using o
mni-directional antennas. Sets of signal attenuation curves were plotted as a fu
nction of frequency and distance by which relevant gain factors were determined.
Okumura calculated that the base station antenna height gain factor varies at a
rate of 20 decibels per decade, and the mobile antenna height gain factor varie
s at a rate of 10 decibels per decade for heights less than three meters. Terrai
n corrections such as undulation height, isolated ridge height, and average slop
e can be applied to the Okumura model. The correction factors are published as p
lotted curves.
The Hata equation model is appropriate if you do not have detailed terrain infor
mation and are working in urban or suburban environments. The Planet Hata equati
on model includes the COST 231 extensions from 1 500 MHz to 2 000 MHz.
The Okumura model performs well for cellular systems in cluttered environments w
ith common standard deviations between predicted and measured path loss values o
f approximately 10 to 14 decibels. Hata has reduced the main results of Okumura
et al. to a few equations, and an application of these equations is commonly kno
wn as the Okumura-Hata method.
Model versions
Two versions of the Okumura-Hata propagation model are shipped with Planet: 2.0
and 2.5. If you are building a new project, you can use version 2.5 of the Okumu
ra-Hata model.
The Hata method requires an average terrain elevation from the transmitter to th
e receiver. Averaging starts at 3 kilometers and goes to the receiver, or to 15
kilometers, whichever is less. If the receiver is less than 3 kilometers away fr
om the transmitter, there is no average; the terrain height at the receiver is u
sed. Version 2.0 of the Okumura-Hata propagation model calculates the average to
15 kilometers in all cases. If you have sites in a valley and have been getting
excessively small predicted signal strengths, you can reconfigure these sites u
sing version 2.5 of the Okumura-Hata model.
Planet General Model
The Planet General Model is a flexible hybrid model that can be used to model ma
ny different kinds of propagation environments. It enables you to migrate data f

rom Planet 2.8 to Planet and obtain the same coverage results as Planet 2.8.
You can use the Planet General Model to model many different kinds of propagatio
n environments. The path loss equation incorporates losses due to a number of mo
dels (such as Okumura-Hata), contributors, and coefficients that can be pieced t
ogether to create a user-defined propagation model. Some of these are defined by
algorithms derived from statistical data. These algorithms are quite accurate u
nder specific conditions, but become less appropriate as the terrain and clutter
varies from these conditions. Various correction factors exist to compensate fo
r these varying conditions, and it is very important for these values to be assi
gned accurately in order to make models simulate the real situation.
The Planet General Model predicts the path loss
iction area. This is achieved by constructing a
the base station (transmitter) to each element
for that profile. In order to ensure that path
prediction region is computed, a profile can be
he perimeter of the prediction region. Thus the

for each element within the pred

terrain and clutter profile from
and then computing the path loss
loss at each element within the
constructed to each element on t
number of radials, , is given by

However, for most practical applications, a fraction of the above number of radi
als is sufficient. A corresponding signal strength at each element is also compu
ted using the antenna pattern.
One of the most visible differences between the Planet General Model used with P
lanet 2.8/Planet DMS and the one used with Planet is the shape of the prediction
area; Planet 2.8/Planet DMS uses a square prediction area, whereas Planet defin
es a circular prediction area. Although the shape and the total area of the pred
iction areas are markedly different, this has no effect on the computed path los
s or signal strength values. Using simple geometry, you can convert Planet 2.8 P
rediction Size to Planet Propagation Distance using
The above equation overlaps the Planet circular prediction area with Planet 2.8
square prediction region, thus assuring total coverage of the prediction zone.
For more information on the Planet General Model, see the Planet General Model T
echnical Note. Application Notes and Technical Notes are available from the Ment
um web page.
ITU 370-Recommendation model
You can use the ITU 370-Recommendation 2.5 propagation model for modeling VHF an
d UHF broadcast services.
Planet includes both the ITU 370-Recommendation 2.0 model and the 2.5 model. Ver
sion 2.0, which lacks the model tuning capability of version 2.5, has been added
for backward compatibility with existing projects. The ITU 370-Recommendation m
odel is the implementation of ITU Recommendation ITU-R P.370-7 and is designed s
pecifically for broadcast services in the VHF and UHF bands. The model is based
on propagation curves and correction factors that determine the dependency of si
gnal strength on transmitting-antenna height and on the distance from a transmit
ter. Each propagation curve shows the effect of the frequency band, landscape ty
pe, and the percentage of time on the signal strength. In Planet, you can specif
y percentages of time and of locations, frequency mode, bandwidth, environmental
settings, and terrain factors. The ITU 370-Recommendation model provides coeffi
cients of correction for Rural, Suburban, and Urban clutter types, which are use
r selectable. You cannot make any numeric adjustments (e.g., dB) to the clutter
attenuation.

The ITU 370-Recommendation model is best suited to frequencies between 30 and 10

00 MHz and distances up to 1000 kilometers.
Interpreting Recommendation 370 results
When you are interpreting Recommendation 370 results, keep in mind the following
points:
The signal strengths in the ITU 370-Recommendation models refer to one kilowatt
Effective Radiated Power (ERP) from a half-wave dipole. However, Planet adjusts
the results to the sector s parameters in the site table.
The basic calculation accommodates for any effective transmitter antenna height,
while the receiving antenna height is fixed at 10 meters. However, a height gai
n function in the ITU 370-Recommendation models allows you to consider other rec
eiving antenna heights.
The land path curves refer to the value of terrain irregularity at 50 meters, wh
ich generally applies to rolling terrain commonly found in Europe and North Amer
ica. The ITU 370-Recommendation models also include a terrain-clearance-angle co
rrection that depends on the terrain close to the receiver.
COST 231 Walfisch-Ikegami model
You can use the Walfisch-Ikegami model for urban or suburban areas with uniform
building heights and separation on flat ground.
COST 231 has proposed a combination of the Walfisch and Ikegami models that has
been accepted by the ITU-R and included in Report 567-4. This model is statistic
al and not deterministic, because terrain and clutter are not considered.
The parameters used by the model are shown in Figure 4.1. When you use the model
, you need to input the height of the buildings (hRoof), the widths of roads (w)
, the building separation (b), and the road orientation. The parameters that you
define in Planet include the transmitter height, the receiver height, and the f
requency.
Figure 4.1 COST 231 Walfisch and Ikegami model parameters
The model distinguishes between line-of-sight (LOS) and non-line-of-sight (NLOS)
situations. The LOS case describes a street canyon situation, such as when the
transmitter is located at a street corner and LOS is achieved in the direction o
f the streets. The NLOS case uses the building and street properties to estimate
the path loss at a given location.
COST 231 has defined the following restrictions on the model:
Frequency: 800-2000 MHz
hBase: 4-50 m
hMobile: 1-3 m
Distance: 0.02-5 km

Planet does not restrict the range of these parameters; therefore, predictions m
ust be considered with care outside of these ranges.

The estimation of path loss agrees rather well with measurements for base statio
n antenna heights above roof-top levels. The error becomes larger when hBase is
approximately equal to hRoof. The performance of the model is quite poor when hB
ase is much less than hRoof.
The parameters b, w, and f are not considered in a meaningful way for microcells
. Therefore, the prediction error in microcells might be quite large.
The model does not consider multipath propagation, and the reliability of the pr
ediction decreases if the terrain is not flat or the clutter is not homogeneous.
Longley-Rice model
You can use the Longley-Rice area calculation for rural (non-urban) areas if lit
tle is known about the terrain and clutter.
The Longley-Rice model is applicable to point-to-point communication systems in
the 20 MHz to 10 GHz range over different types of terrain (Rappaport, 1996). Th
e Longley-Rice model operates in two modes. The point-to-point mode uses terrain
information if it is available, while the point-to-area mode uses techniques th
at estimate the path-specific parameters when little terrain information is avai
lable.
In point-to-point mode, median path loss is predicted by using tropospheric refr
activity and terrain geometry. However, only some features of the terrain are us
ed. The terrain profile is used to find effective antenna heights, horizon dista
nces and elevation angles as seen from the antennas, the angular distance for a
trans-horizon path, and the terrain irregularity of the path. The prediction is
performed in terms of these parameters. A ray optic technique using primarily a
two-ray ground reflection model is used within the radio horizon. The two or thr
ee isolated obstacles causing the greatest obstruction are modeled as knife edge
s using the Fresnel Kirchoff theory. Forward scatter theory is used to make trop
oscatter predictions for long paths and far field diffraction losses are predict
ed using a modified Van der Pol-Bremmer method (Rappaport, 1996). The Longley-Ri
ce point-to-point model is also referred to as the Irregular Terrain Model (ITM)
(Hufford, et al. 1982).
Although the point-to-area mode is an old method, it is still perhaps the best m
ethod of estimating path loss in open country if the only parameters known about
the ground are its irregularity and (less importantly at UHF) its electrical co
nstants.
The Longley-Rice model is best suited to the following parameters:
Frequency: 20 MHz to 10 GHz
Distance: 1 km to 2000 km
Antenna Heights: 0.5 m to 3000 m
Polarization: Vertical or Horizontal
References
For more information about the Longley-Rice model, see the following references:
Rappaport, T.S. Wireless Communications: Principles and Practice. Prentice Hall,
1996.

Hufford, Longley, and Kissick. A Guide to the Use of the ITS Irregular Terrain Mo
del in the Area Prediction Mode , U.S. Department of Commerce. April 1982.
Lee model
You can use the Lee propagation model when you have survey results that show the
nature of signal decay for local propagation conditions. The Lee model combines
both an analytical and experimental approach to the estimation of both signal s
trength and path loss.
The standard equation for the Lee propagation model is described below.
Where:
is the mean received signal level at distance R from the transmit antenna.
is the expected signal strength in dBm for the reference conditions defined by
, , , and .
is the slope or rate of signal strength decay as a function of distance from th
e transmitter in dB/decade.
is the distance from the transmitter in kilometers.
is the reference distance from the transmitter in kilometers.
is the effective antenna height of the transmitter in meters.
is the antenna height of the reference transmitter in meters.
is the effective antenna height of the receiver in meters.
is the antenna height of the reference receiver in meters.
is the effective radiated power of the transmitter in watts.
is the effective radiated power of the reference transmitter in watts.
is the knife-edge diffraction losses or additional loss due to terrain obstruct
ion.
is the antenna pattern gain or additional loss or gain as a result of the actua
l antenna pattern used in the prediction.
The Lee model relies on a set of path loss curves that apply to a reference tran
smitter. These curves are straight lines on a logarithmic scale of distance, and
are defined by a slope (a) and an intercept at 1.0 or 1.6 kilometers. These par
ameters are usually obtained from survey measurements that show the speed of sig
nal decay as a function of distance under local propagation conditions. The Lee
model formula calculates the signal strength at any given point by modifying the
reference signal strength to take into account the distance, the antenna height
s, and so on actually encountered.
If the terrain is flat, nothing more is done. With hilly terrain, the terrain da
ta is used to calculate an effective antenna height for the transmitting antenna
, and also to estimate the additional path loss due to terrain obstructions mode
led as knife edges. The changes in signal strength due to a modified effective a
ntenna height and due to the knife-edge obstructions are added to the signal str
ength calculated for flat terrain.

IEEE 802.16 model

You can use the IEEE 802.16 model when designing Multipoint Distribution System
(MDS) and LMDS networks with frequencies in the 10 to 66 GHz range. This frequen
cy range is characterized by very high data rates and short range due to rain an
d foliage attenuation.
The IEEE 802.16 model is recommended for use with broadband wireless access tech
nologies.
Terrain types
The following types of terrain are recommended for use with the IEEE 802.16 mode
l:
Type A characterized by hilly, moderate-to-heavy tree density (for light to modera
te urban areas)
Type B characterized by hilly, light tree density or flat, moderate-to-heavy densi
ty
Type C characterized by flat, light tree density
Path loss equation
The standard path loss equation for the IEEE 802.16 model is described below.
The path loss calculation only accounts for the following parameters:
transmitter height
receiver height
frequency
the ground type as defined in the IEEE.802.16 dialog box
The clutter grid, the elevation file, the rain attenuation, and clutter absorpti
on losses have no effect on the path loss calculation.

Where:
is equal to .
is the wavelength in meters.
is the path loss exponent equal to .
is the height of the base station in meters.
is equal to 100 m.
, , and are constants dependent on the terrain type.
is a statistical term for random shadow fading (zero mean).

References
For more information about the IEEE 802.16 model, see the following references:
Erceg, Vinko, et al. An Empirically Based Path Loss Model for Wireless Channels i
n Suburban Environments . IEEE Journal on Selected Areas in Communications. Vol. 1
7, No 7, July 1999.
The IEEE 802.16 Working Group on Broadband Wireless Access Standards web site at
http://ieee802.org/16.
Chang, D.K. IEEE 802.16 Technical Backgrounder . IEEE 802.16 Broadband Wireless Acc
ess Working Group. May 2002.
CRC-Predict model
You can obtain information about CRC-Predict model properties by pressing the F1
key from the Predict Parameters or the Predict Properties dialog box. For more
detailed information, see the CRC-Predict Propagation Model Technical Note. Plan
et Application Notes and Technical Notes are available from the Mentum web page.
CRC-Predict is a general-purpose model intended for macrocell planning. It is no
t a ray-tracing model and, as such, should not be used with high-resolution data
. Instead, it is best used with geodata with a resolution between 20 to 30 meter
s. You can use it in most circumstances, regardless of the kind of terrain, if d
etailed terrain or clutter information or both are available. The following case
s are exceptions:
for very short paths, for example micro-cellular paths, in which the locations o
f individual buildings are important
when a very rapid calculation is wanted, because the CRC-Predict model is more c
omputationally intensive than most models
The path loss calculation in the CRC-Predict model is designed for the VHF to UH
F (30 MHz to 3 GHz) frequency range. The physical principles used by the CRC-Pre
dict model are also applicable up to 30 GHz. However, accurate predictions for t
hat range depend on very detailed and accurate terrain data, and currently there
are no supporting test measurements. Also, above 10 GHz, rain attenuation becom
es significant. The principal algorithm is a diffraction calculation, based on t
he Fresnel-Kirchoff theory that takes terrain into account in a detailed way. An
estimate of the additional loss for obstructions such as trees, buildings, or o
ther objects is included when data on clutter classes are available. Tropospheri
c scatter is included for long paths. Estimates of time and location variability
can be made.
The diffraction algorithm samples the propagation path from the transmitter to t
he receiver and determines the signal strength at many points in space. First, t
he wave field is determined as a function of height (a vertical column of many v
alues) above a terrain point close to the transmitter by an elementary calculati
on. Then, using the Huygens principle of physical optics, each of these field po
ints is regarded as a source of radiation, and from them, the signal strength is
calculated a little farther away. In this way, a marching algorithm simulates t
he progress of the radio wave from the transmitter to the end of the path. Even
though the signal strength is calculated at many points, an efficient integratio
n algorithm and a choice of only the most important signal strength points permi
t the integration calculation to be fast enough for practical use.
The CRC-Predict model also uses surface-type or clutter data in its calculations
. See Appendix C: Clutter Properties on page 443. Because CRC-Predict is a determi

nistic model, the more precise and physically realistic terrain and clutter info
rmation you use, the more accurate the output tuned model will be.
Clutter interacts with the algorithm in two ways:
As the wave propagates over the ground toward a distant receiver, the effective
height of the ground is assumed to be the real height of the ground plus the ass
umed clutter height.
Clutter close to the receiver is assumed to terminate close to the receiver, e.g
., 50 meters. That is, the receiving antenna is not assumed to be on the doorste
p of a building, or in the middle of a forest, but rather on a street or in a ro
ad allowance in the forest. Part of the calculation is an estimate of the attenu
ation from the clutter down to street level.
In addition to the height and distance of solid (opaque) clutter, there is an ad
ditional attenuation, entirely empirical, which takes into account trees and oth
er absorbing material adjacent to the receiving antenna. This attenuation factor
(expressed in decibels) is the parameter most easily used to make median predic
tions agree with measurements in a particular area (model tuning).
CRC Predict-Air
Only masked path loss is calculated and saved in the prediction files. As a resu
lt, if you change any site setting (other than transmitted power), all of the pr
ediction files are regenerated.

CRC-Predict Air is a unique model designed for high-altitude communication (e.g.

, aircraft to ground) where the signal is being broadcast upwards (between 0 and
+90/-90 degrees). It is based on the CRC-Predict 4.0 propagation model. You can
use the CRC-Predict Air model in two modes:
AMSL (Above Mean Sea Level) mode in this mode, you can define the antenna height.
For the purpose of propagation calculation, the receive height remains at a con
stant height above sea level.
AGL (Above Ground Level) mode in this mode, the receiver antenna height will be re
lative to the ground level as defined by the input Digital Elevation Model (DEM)
.
Unlike the CRC-Predict model, this new model will not generate path loss predict
ions (grid files) which can be re-masked. It is also important to note that you
cannot tune CRC-Predict Air models.
The Point-to-Point tool does not support the CRC-Predict Air propagation model;
however, the CRC-Predict 4 model provides results similar to the CRC-Predict Air
model when used in AGL mode.
Universal model
The Universal model is only available if you have purchased a license. You can o
btain detailed information about the Universal model by pressing the F1 key from
the Universal Model Parameters dialog box. The online Help for this model conta
ins context-sensitive help, as well as the Universal Model User Guide.
The Universal model is a high-performance deterministic propagation model that h
as been integrated into Planet. Unlike other propagation models, the Universal m
odel automatically adapts to all engineering technologies (i.e., micro, mini, sm

all, and macro cells), to all environments (i.e., dense urban, urban, suburban,
mountainous, maritime, and open), and to all systems (i.e., GSM, GPRS, EDGE, UMT
S, WIFI, and WIMAX) in a frequency range that spans from 400MHz to 5GHz.
In addition, the Universal Model:
uses a new AGL layer and a new polygon layer where modifications to the layers c
an be done directly in the Map window.
outperforms other models in terms of the speed and accuracy of predictions.
Q9 model
The Q9 propagation model is based on the Okumura-Hata model. Using the variables
shown in Figure 4.2, it calculates the expected pathloss between the transmitte
r and the receiver using the terrain profile. In other words, it considers a cro
ss-section of the earth along a straight line between the transmitter and the re
ceiver. This propagation model is most useful for frequency bands in the 150-200
0 MHz range and works best within a radius of 0.2-100 km. The Q9 model is intend
ed for use with high-resolution elevation and clutter data.
Pathloss depends on frequency as well as the antenna heights of the transmitter
and the receiver. The Q9 model allows for both uptilt and downtilt of antennas a
nd takes into account the vertical antenna pattern.
There are three input values that the Q9 model considers:
Okumura-Hata s wave propagation equations with modifying parameters A0 to A3. See
Equation 4.2 on page 155. For more information on the A0 to A3 parameters, press
the F1 key in the Q9 Parameters dialog box.
Extra losses that occur when wave propagation is disturbed by obstacles such as
mountain peaks. When the distance between the transmitter and receiver becomes s
ufficiently large, a correction due to earth s curvature is necessary.
Land use code loss.
The Figure below illustrates the variables that are taken into account to calcul
ate pathloss.

Equation 4.1 details the formula used to calculate pathloss.

Equation 4.1 Pathloss calculation

Where:
Lb is the pathloss
HOA (Hata Open Area) is a variant of Okumura-Hata s equation in dB as shown in equ
ation Equation 4.2
mk[mobile] is the land use code at the mobile in dB
is a parameter related to the knife-edge diffraction
KDFR is the contribution from knife-edge diffraction in dB

JDFR is the diffraction loss due to the spherical earth in dB

Equation 4.2 Hata Open Area equation

Where:
A11 is equal to A1 x log d
g(F) is equal to 44.49 x logF - 4.78 x (logF)2
HEBK is the effective antenna height in meters as defined in the Q9 propagation
model.
d is the distance from the base antenna to the mobile in kilometers
A0, A1, A2, A3 are Q9 model tuning parameters

WaveSight model
The WaveSight model is only available if you have purchased a license. You can ob
tain detailed information about the WaveSight model by pressing the F1 key from
the WaveSight Model Properties dialog box. The online Help for these models cont
ains context-sensitive help, as well as the WaveSight User Guide.
The WaveSight model is based on the uniform theory of diffraction. To predict th
e signal power, the WaveSight model takes individual buildings and vegetation, a
s well as terrain and clutter, into account.
The WaveSight model is not restricted to specific environments. It can be applie
d in urban, suburban, rural, and open areas. However, most of the tests on the m
odel were conducted in urban and suburban areas. No tests were conducted for a r
adius greater than 20 km.
Because of the physical nature of the model, which uses the uniform theory of di
ffraction, frequency is a parameter of the model. Extensive tests were performed
in the 800, 900, 1800, and 2000 MHz bands.
The WaveSight model enables computations with no limitation on transmitter or re
ceiver heights; however, no drive test data was available for receiver heights g
reater than 2 m above ground.
The WaveSight model uses raster data, e.g., terrain and clutter, in a format sim
ilar to that used with Planet DMS. In certain cases, the raster data is availabl
e in several resolutions typically a resolution of 20 m or more for a large area s
uch as an entire state or nation, and 5 m for small built-up areas. In such case
s, the WaveSight model uses the highest available resolution associated with the
area under consideration.
The required accuracy is 2 m on the wall position. All buildings with a footprin
t larger than 16 m2 must be represented in the building database. The WaveSight
User Guide lists the consistency rules required from the vector database, i.e.,
no open polygons or building overlap.

One of the input parameters used by the WaveSight model is the attenuation loss
incurred going from outdoor to indoor. The WaveSight model uses this value to co
mpute the signal strengths inside the building.
Wavecall is constantly improving WaveSight performance on an increasing pool of
measurements. Whenever a divergence between model and data is observed, the mode
l is updated and retested on all available routes to ensure that the modified mo
del is consistent with experimental data. Therefore the overall performance of t
he model is constantly increased. Thus, in general, there is no need for the mod
el to be tuned.
Because of the subjective nature of the clutter, tuning is advisable in open and
rural areas where clutter significantly influences propagation. Tuning must be
applied with care and only when there are sufficient measurement samples availab
le that are representative of the environment.

Did you find this topic helpful?