PROPAGATION

MODELS AND
MODEL TUNING

Svemir Polchert
PROPAGATION
ElectRomagnetic Waves
Propagation
› The mechanism behind electromagnetic waves
propagation are diverse, but can generally be divided into
four types ：
› Direct signal
› Reflection Direct Signal
Reflected Signal
› Diffraction
Transmitter
› Scattering
Scatted Signal Receiver
Diffracted Signal

EUS/WE/CM Svemir Polchert | Ericsson Confidential | EUS-13:008212 Uae, Rev A | 2013-07-15 | Page 3
Mobile Radio Channel
Characterized by
›Path Loss
›Slow Fading
›Fast Fading
Received Power Fast fading =
Rayleigh or Ricean
Distribution

Slow fading =
Log-normal
Distribution

Distance
EUS/WE/CM Svemir Polchert | Ericsson Confidential | EUS-13:008212 Uae, Rev A | 2013-07-15 | Page 4
PROPAGATION
MODELS
General PROPAGATION
Model Assumptions
› Path loss with distance and frequency

› Path loss with increasing base station and mobile

antenna heights

› Path loss is affected by clutter type, diffraction, weather

conditions, time of the year, proximity of other obstacles and
street orientation

EUS/WE/CM Svemir Polchert | Ericsson Confidential | EUS-13:008212 Uae, Rev A | 2013-07-15 | Page 6
General PROPAGATION
Theory
› Free space model

› Plane Earth Propagation - factor for reflections off plane earth

› Egli model - Plane Earth Propagation + frequency dependent
factor
› Knife edge diffraction - Fresnel parameter to cater for single
knife-edge

EUS/WE/CM Svemir Polchert | Ericsson Confidential | EUS-13:008212 Uae, Rev A | 2013-07-15 | Page 7
General Theory -
Okumura
› Okumura - empirically based model made from
measurements in Tokyo, 1968
› Frequency and mobile height dependent
› Model is seen as a series of curves
› Correction factors for open area, suburban, sloping terrain
etc.
› Validity:
–f : 150 - 2000 MHz
– Dist: 1 - 100 km
– Station ht: 30 - 1000 m

EUS/WE/CM Svemir Polchert | Ericsson Confidential | EUS-13:008212 Uae, Rev A | 2013-07-15 | Page 8
General Theory - Hata

› Empirical formulas based on Okumura’s model for computational

use
› Path Loss:

› Correction factors for urban, suburban, open areas.

› Validity:
–f : 150 to 1500 MHz
– Dist: 1 to 20 km
– Station ht: 30 to 200 m
– Mobile ht: 1 to 10 m

EUS/WE/CM Svemir Polchert | Ericsson Confidential | EUS-13:008212 Uae, Rev A | 2013-07-15 | Page 9
 Propagation Model TYPES
Increasing

› No single model is Accuracy

suitable for every

application
– Model selection
depends primarily 50
on the application, m
type of GIS data and CRC Predict Planet General Model or
type of drive test
(Deterministic) Atoll Standard Model
data
Resolution
GIS Data

› Deterministic model 25
(Semi-Empirical)
simulates behavior of m
the radio wave as it
Universal Model (Planet) = Crosswave (Atoll)
encounters terrain
features or buildings (Deterministic & Semi-Empirical)
Ray-tracing (Deterministic)
› Semi-empirical model 5
m No Hype
utilizes measurement Few Many
r
data to define a model Tunin Model
Tuning
Model
path loss equation g s s Tunin
Mode g

EUS/WE/CM Svemir Polchert | Ericsson Confidential | EUS-13:008212 Uae, Rev A | 2013-07-15 | Page 10
MODEL TUNING
Propagation Model Tuning -
Why and When
› Fundamental aim to cell planning - build
sites to meet specific coverage and
capacity objectives
› Tuned model allows for more accurate 0 dB

estimation of site coverage Theoretically Ideal

› Choose model based on your Model

Percentage of Drive Test Points

requirements
Tuned Model
› Model Tuning is performed at any point in
time when accurate predictions (for site
coverage) is required. This can be: Un-tuned
Model
– in Network Rollout after contractual signing
– in Early Start projects (pre-contract)
– in live networks - Drive Test minus Prediction
+

EUS/WE/CM Svemir Polchert | Ericsson Confidential | EUS-13:008212 Uae, Rev A | 2013-07-15 | Page 12
Geodata
› Terrain and Clutter is the minimum requirement for non Ray-tracing models
› Clutter Heights geodata can be used to enhance accuracy of non Ray-tracing models
› For Ray-tracing models Heights, Clutter and Building Polygons are the minimum
requirement
Clutter Heights

Clutter

Heights

EUS/WE/CM Svemir Polchert | Ericsson Confidential | EUS-13:008212 Uae, Rev A | 2013-07-15 | Page 13
Building Polygons
Drive Test Data

› CW is suitable for any propagation model type

› Scanner data
– Not suitable for Planet General Model or Atoll Standard Model because it
doesn’t capture slope component accurately
– Suitable for CRC Predict and Ray-tracing models

EUS/WE/CM Svemir Polchert | Ericsson Confidential | EUS-13:008212 Uae, Rev A | 2013-07-15 | Page 14
Proven MODEL TUNING
Methodology (1 of 2)
Step 1- Project Requirements Development
›Prioritize areas to analyze with special attention to areas lacking drive test data correlation
›Process inputs, clutter data review, terrain variations and consideration of location specific effects are
examined in detail
›Determine most suitable market modeling strategy

Step 2- Selection of a Suitable Model(s)

›Appropriate type and kind of model depends on
– Propagation tool
– RF environment to model
– Required accuracy
– Available GIS data
– Amount and type of drive test data the drive test budget can support

Step 3 - Development of Data Collection Methods

›Best representative sites are selected
›Segmented by zones indicating the best test drive path

EUS/WE/CM Svemir Polchert | Ericsson Confidential | EUS-13:008212 Uae, Rev A | 2013-07-15 | Page 15
Proven MODEL TUNING
Methodology (2 of 2)
Step 4 - Data Collection (Drive Testing)
›Actual signal strength is measured along pre-determined routes around the sites

Step 5 - Drive Test Post-processing

›Drive test data is averaged to mitigate fast fading
›Filter out possible drive test errors introduced by the test hardware limitations, uncharacterized
structures like over-passes and tunnels and data influenced by larger near-field obstructions

Step 6 - Model Tuning

›Models are tuned by utilizing Model Tuning Tool in the propagation prediction tool

Step 7 - Model Validation

›Calibrated models are tested and evaluated by RF engineers
– Rigorous acceptance process

EUS/WE/CM Svemir Polchert | Ericsson Confidential | EUS-13:008212 Uae, Rev A | 2013-07-15 | Page 16
Model Tuning Results
› Mean error, Standard deviation, RMS and Correlation
factor are used to validate the performance of Tuned Model
Name of the survey Mean Error Standard deviation Correlation factor
1900_ILU0290_ILU02907_filtered -0.4449 8.3541 0.5478
1900_ILU1226_ILU12265_filtered -0.1684 9.0816 0.5717
1900_ILU1226_ILU12266_filtered -3.0471 8.1741 0.8148
1900_ILU1227_ILU12277_filtered 1.8543 5.5329 0.8745
1900_ILU1227_ILU12278_filtered 1.9655 6.7229 0.8225
1900_ILU1227_ILU12279_filtered 0.735 8.3101 0.8014
1900_ILU1308_ILU13087_filtered 0.3416 8.2238 0.8609
1900_ILU3228_ILU32289_filtered -0.2763 9.2539 0.7679
1900_ILU3240_ILU32408_filtered 0.5141 7.3019 0.7342
1900_ILU3240_ILU32409_filtered 0.8567 7.6552 0.7978
1900_ILU3290_ILU32909_filtered -3.092 7.9015 0.7209
1900_ILU3621_ILU36219_filtered 0.3993 7.2545 0.7747
1900_ILU3973_ILU39738_filtered -7.221 8.6051 0.7384
1900_1_ILU1224_ILU12246_filtered -1.7856 7.5029 0.6398
1900_1_ILU1233_ILU12336_filtered 0.4528 7.2815 0.6728
1900_1_ILU1235_ILU12358_filtered 5.7771 8.3443 0.4875
1900_1_ILU1613_ILU16138_filtered -3.1498 8.8738 0.6641
1900_1_ILU1613_ILU16139_filtered 1.6892 7.5266 0.704
1900_1_ILU3181_ILU31818_filtered -0.4034 7.073 0.8625
1900_1_ILU3311_ILU33118_filtered 2.6265 7.3218 0.8293
1900_1_ILU3952_ILU39527_filtered 1.3982 9.78 0.7323
1900_1_ILU3952_ILU39528_filtered 3.2914 8.761 0.6921
1900_2_ILU3308_ILU33088_filtered 2.6368 8.4431 0.7862
1900_2_ILU3621_ILU36218_filtered -3.5616 7.2055 0.7045
1900_2_ILU4332_ILU43327_filtered 2.1228 8.7188 0.6554
Global Performances 0.2536 8.2931 0.8063
EUS/WE/CM Svemir Polchert | Ericsson Confidential | EUS-13:008212 Uae, Rev A | 2013-07-15 | Page 17
SURVEY VS. PREDICTION
UM MODEL UNTUNED UM MODEL TUNED

EUS/WE/CM Svemir Polchert | Ericsson Confidential | EUS-13:008212 Uae, Rev A | 2013-07-15 | Page 18
›
Virtual Site Validation
For Ray-tracing models propagation tool needs to have antennas placed as
close to real antenna location in order to make prediction accurate
› Google or Bing maps are used to find the antenna locations
– Only location of the antenna locations can be verified
› Antennas need to be also moved outside the polygons if the antenna height is
same or lower then polygon height

EUS/WE/CM Svemir Polchert | Ericsson Confidential | EUS-13:008212 Uae, Rev A | 2013-07-15 | Page 19
 IN-BUILDING PROPAGATION
IN UM / Crosswave
› User-defined penetration parameter
allows activation of the outdoor to
indoor process
› Outside->Inside parameter sets In-
building penetration loss in dB to the
receivers situated inside a building
› Inside->Inside parameter sets the
value in dB/m of the assigned
additional loss receivers inside a
building

EUS/WE/CM Svemir Polchert | Ericsson Confidential | EUS-13:008212 Uae, Rev A | 2013-07-15 | Page 20