Model Tuning Report

NOKIA RNP Team

PT.Nokia Solutions and Networks, April 2015

Revision History
Date

Revision No.

Description

Author

Contents
1

Introduction ........................................................................................................... 6

1.1

Model Tuning Process ............................................................................................. 6

1.2

Model Tuning Objectives ......................................................................................... 6

1.3

Digital Map .............................................................................................................. 6

1.4

Propagation Model Definition................................................................................... 8

1.4.1

Propagation environments ....................................................................................... 8

1.4.2

Standard Propagation Model ................................................................................... 8

Site Selection and Route Planning Process ........................................................ 10

2.1

RF Measurement..................................................................................................... 10

2.2

Site Selection .......................................................................................................... 10

2.3

Site survey .............................................................................................................. 12

2.4

CW Test route ......................................................................................................... 12

Model Calibration .................................................................................................. 13

3.1

Introduce ................................................................................................................. 13

3.2

Model for Dense Urban ........................................................................................... 13

3.2.1

Analysis CW Test Data for Dense Urban ................................................................. 13

3.2.2

Model Tuning Result for Dense Urban ..................................................................... 23

3.3

Model for Urban....................................................................................................... 24

3.3.1

Analysis CW Test Data for Urban ............................................................................ 24

3.3.2

Model Tuning Result for Urban ................................................................................ 34

3.4

Model for Sub Urban ............................................................................................... 35

3.4.1

Analysis CW Test Data for Sub Urban ..................................................................... 35

3.4.2

Model Tuning Result for Sub Urban......................................................................... 43

3.5

Model for Rural ........................................................................................................ 44

3.5.1

Analysis CW Test Data for Rural ............................................................................. 44

3.5.2

Model Tuning Result for Rural ................................................................................. 50

4. LTE Propagation Model of Jakarta................................................................................. 51

Figures
Figure 1-1 clutter plot of Jakarta ................................................................................................... 7
Figure 2-1 Rooftop antenna clearance ....................................................................................... 11
Figure 3-1 CW Signal Strength plot Dense Urban ...................................................................... 13
Figure 3-2 signal Strength vs 10 log (d) before filter UOB .......................................................... 14
Figure 3-3 signal Strength vs 10 log (d) before filter Mandiri ...................................................... 15
Figure 3-4 signal Strength vs 10 log (d) before filter PPHUI ....................................................... 16
Figure 3-5 signal Strength vs 10 log (d) before filter Resto Paregu ............................................ 17
Figure 3-6 Dense Urban DT Data process ................................................................................. 18
Figure 3-7 Signal strength vs log (d) after filter UOB .................................................................. 19
Figure 3-8 Signal strength vs log (d) after filter Mandiri .............................................................. 20
Figure 3-9 Signal strength vs log (d) after filter PPHUI ............................................................... 21
Figure 3-10`Signal strength vs log (d) after filter Resto Paregu .................................................. 22
Figure 3-11 CW Signal Strength plot Urban ............................................................................... 24
Figure 3-12 signal Strength vs 10 log (d) before filter Salon JLO................................................ 25
Figure 3-13 signal Strength vs 10 log (d) before filter Kemang Selatan ...................................... 26
Figure 3-14 signal Strength vs 10 log (d) before filter Cementaid ............................................... 27
Figure 3-15 signal Strength vs 10 log (d) before filter Ruko Artha KEdoya ................................. 28
Figure 3-16 Urban DT Data process .......................................................................................... 29
Figure 3-17 Signal strength vs log (d) after filter Salon JLO ....................................................... 30
Figure 3-18 Signal strength vs log (d) after filter Kemang Selatan .............................................. 31
Figure 3-19 Signal strength vs log (d) after filter Cementaid ....................................................... 32
Figure 3-20 Signal strength vs log (d) after filter Ruko Artha Kedoya ......................................... 33
Figure 3-21 CW Signal Strength plot Sub Urban ........................................................................ 35
Figure 3-22 signal Strength vs 10 log (d) before filter Primagama .............................................. 36
Figure 3-23 signal Strength vs 10 log (d) before filter Bina Asih ................................................. 37
Figure 3-24 signal Strength vs 10 log (d) before filter Perkutut ................................................... 38
Figure 3-25 Sub Urban DT Data process ................................................................................... 39
Figure 3-26 Signal strength vs log (d) after filter Primagama ...................................................... 40
Figure 3-27 Signal strength vs log (d) after filter Bina Asih ......................................................... 41
Figure 3-28 Signal strength vs log (d) after filter Perkutut ........................................................... 42
Figure 3-29 CW Signal Strength plot Rural ................................................................................ 44
Figure 3-30 signal Strength vs 10 log (d) before Rangon Jaya ................................................... 45
Figure 3-31 signal Strength vs 10 log (d) before filter Sukatani .................................................. 46
Figure 3-32 Rural DT Data process............................................................................................ 47
Figure 3-33 Signal strength vs log (d) after filter Rangon Jaya ................................................... 48
Figure 3-34 Signal strength vs log (d) after filter Sukatani .......................................................... 49

Tables
Table 1-1 Clutter Classification scheme ....................................................................................... 8
Table 2-1 List of selected CW test for Jakarta ............................................................................ 11
Table 3-1Test Point count of each site Dense urban .................................................................. 14
Table 3-2Test Point count of each site Urban............................................................................. 24
Table 3-3Test Point count of each site Sub Urban ..................................................................... 35
Table 3-4Test Point count of each site Rural .............................................................................. 44
Table 4-1 Final Propagation Model tuning result of Jakarta ........................................................ 51

1 Introduction
Propagation model Tuning is a crucial procedure early in network Deployment as it enables
accurate predictions of coverage and interference. The objective of this document is to describe
the CW drive campaign and the propagation model tuning procedure for smartfren FDD 850
project in Jakarta of Indonesia. Here we do CW test and propagation model tuning base on FDD
850 MHz, and for LTE propagation model was calculated depend on tuned Standard Propagation
Model.
Anite Nemo Scanner and Coyote used to data process and Atoll (version : 3.2.1.7090) for model
tuning Maps & Clutters
1.1

Model Tuning Process

The model tuning process including the following activities:

Define models required to simulate coverage of different physical environments

Site survey to find suitable sites for each environment
Clearance Scanning
Equipment and testing
CW test and data collection
Assessment and preparation data
Model tuning for each environment
Validate the model

1.2 Model Tuning Objectives

The propagation models included in this investigation refers to dense urban, urban, sub urban,
and rural area.
The following were used as objectives throughout the process:
Mean error between -1 and 1 dB in global calculation each clutter
Standard Deviation error <= 8 dB in global calculation each clutter

1.3 Digital Map

Mapping Data is an integral component in prediction calculations and thus accurate data is
essential for any prediction model to function correctly. It is important that mapping data supplied
is as up to date as possible so that it reflect any recent changes in land use.

Digital Terrain Map

All measurement data was resolved using a 20 m resolution Digital Terrain Map. The vintage of
the map used was Indonesia: West Java, Map 20m, January 15th 2015 from Computa Maps
Company. Universal transverse Mercator co-ordinate system with zone 48s and datum WGS84
is used
Clutter Database
The clutter data classes used for the Jakarta model tuning campaign are listed in was
Indonesia: West Java, Map 20m, January 15th 2015 from Computa Maps Company, with the
clutter distribution shown in figure1.1. The 20 m resolution map was used to represent the land
cover.

Figure 1-1 clutter plot of Jakarta

This data formed a part of the procured maps. The following table shows various classifications
that were defined.
Code

Name
Default Values

Sea

inland water

Wetland

Barren

low vergetation

sparse forest

Forest

Village

residential with trees

10

residential with few trees

11

dense residential

12

Urban

13

dense urban

14

high buildings

15

building blocks

16

comercial/industrial

17

Airport

18
open in urban
Table 1-1 Clutter Classification scheme

1.4

Propagation Model Definition

We use standard propagation model in Atoll for FDD LTE 850 MHz network. The Standard
Propagation Model (SPM) is based on the cost -231 formulas and is suited for predictions in the
850 to 3500 MHz band over long distances (from one to 20 km). It is the best suited to UMTS
and LTE radio technologies
1.4.1

Propagation environments

In Jakarta, the main propagation environment is a mix of dense urban, urban,sub urban and rural
clutter. Considering the planning issues (also height and clutter data of digital map), it is
acceptable to develop three models for the whole city, dense urban, urban, sub urban and rural.
The propagation environment is mostly characterised as residential and commercial buildings
throughout the whole city.

1.4.2

Standard Propagation Model

Below is the Standard Propagation Model used in Atoll.

It consists of parts:
-

General parameter

The basic path loss model

Calculation of the base station effective antenna height

Diffraction Clutter corrections

2 Site Selection and Route Planning Process

2.1 RF Measurement

In order to calibrate the propagation model, a comparison was carried out between the predicted
propagation and actual measured data. This measured data was collected by way of a series of
Continuous Wave (CW) propagation surveys. In these drive test a calibrated test transmitter was
set up at a base station located and received signal strength measurements were made with
along a predetermined drive route
The accuracy of the model is directly related to the validity and accuracy of the CW data

2.2 Site Selection

To ensure validity of the calibration process it was essential that site was selected carefully and
that various parameters required in the calibration process were verified. The site morphology is
also a major factor in determine the extent of survey regarding direction around the site.
Site selection factors include:

Test site measured were representative of typical BTS sites, considering issues such as
the general environment and antenna height surrounding clutter characteristics etc.
They were located in and around the area where the prediction model is to be used so as
to capture a good representation of data in that propagation location thus ensuring a
valid model for that propagation classification.
Rooftop sites were chosen with flat roofs and power outlets (possibly sites with BTS
equipment on the roof) so that test masts and equipment could be installed.
.selected sites height was representative of the relative radiation height of the network.
As this drive campaign was concerned with the calibration of macro cell model, micro
cellular sites or umbrella sites were not be taken in to account.
Selected sites coverage was chosen to minimise anomalous local propagation
phenomena, such as near obstacles shadowing, canyon effect, measurement faults,
etc.

It was ensured that minimum height of the transmit antenna was such that the 3 dB
vertical beam width was free of obstruction from the building. This can be seen in:
H=d tan()
Where =0.5 * (3 dB vertical beam width) + safety margin
A margin of 10 degrees was included

Figure 2-1 Rooftop antenna clearance

List of selected CW Test sites for Jakarta model is shown as the following table

Site Name

Morphologies

Longitude

Latitude

Antenna
Height

ZTE_0219 UOB

Dense urban

106.823

6.19796

38

ZTE_0215 MANDIRI

Dense urban

106.8152

-6.225

24

ZTE_0213 PPHUI

Dense urban

106.8328

-6.222

25

ZTE_0206 RESTORAN PAREGU

Dense urban

106.83

-6.1874

32

ZTE_0022 SALON J LO

Urban

106.9009

-6.1575

23

ZTE_0349 CEMENTAID

Urban

106.905

-6.2035

30
23

ZTE_0564 RUKO ARTHA

KEDOYA

Urban

106.759

6.17509

ZTE_0083 KEMANG SELATAN

Urban

106.816

-6.275

42

ZTE_3378 BINA ASIH

Sub Urban

106.9577

-6.3047

42

Sub Urban

106.729

-6.3424

25

Sub Urban

106.6224

-6.141

29

Rural

107.179

-6.1679

60

107.366

6.28468

55

ZTE_0116 PRIMAGAMA
PAMULANG
ZTE_4272 PEKUTUT
TANGERANG
ZTE_2057 SUKATANI
ZTE_2029 RANGON JAYA

Rural

Table 2-1 List of selected CW test for Jakarta

EIRP
(dBm)
53
53
53
47
53
47
53
47
53
47
47
53
53

Antenna
Power
(dBi)

Antenna type

11

K 736347

11

K 736347

11

K 736347

11

K 736347

11

K 736347

11

K 736347

11

K 736347

11

K 736347

11

K 736347

11

K 736347

11

K 736347

11

K 736347

11

K 736347

2.3

Site survey

Factor for planning surveys are followings:

Drive test must first be planned according to the limitations observed through the site survey.

2.4

It is important to collect a statistically significant amount of data to model diffraction. A

good balance between LoS and NLoS should be attempted.
The data should be evenly distributed with respect to distance from the transmitter.
Distance should also be taken into account on a per clutter type basis. Through using
various test site locations this is practically implemented.
When planning drive test routes it is of great importance to ensure that the drive goes
through the clutter type in mind, since consecutive roads may be classified as different
clutter types. If in sufficient data measurements are collected in particular morphology
class in the calibration toward other classes may occur.
Using partially the same routes from different sites is beneficial since the different location
of the test antenna will provide different data with respect to the distance.
The extent of the survey is dictated on the amount of clutter types and required bins along
with the actual purposes of the survey data. In cases where the data collected will be
used for analyzing interference between sites survey may tend to reach long distances
away from the site (up to 20 km) with the actual route exceeding 100 km.
There should be sufficient data collected within each clutter category to ensure accurate
modelling.
CW Test route

The following factors should be considered when planning a route:

The route should be planned according to limitations noted at the survey stage. If the
antenna blocked in any direction then the route should avoid the area affected by this
blocking
The accuracy of the model calibration is dependent upon the amount of data collected, so
the route should cover as much road as timescale permit
The data for each clutter type should be evenly distributed with respect to log (distance)
between the measurement equipment and transmitter
Using partially the same routes for different surveys is beneficial since the different
location of the test antenna will provide different data with respect to distance and
effective antenna height.
The extent of the survey is dictated on the number of clutter types and required
measurements along with the actual purposes of the survey data and the frequency being
used.
The route should incorporate a variety of different terrain variations
Both line of sight and non line of sight points should be covered

3 Model Calibration
3.1 Introduce

For output accurate propagation model for each clutter environment, the model calibration uses
few sites to do propagation model tuning in each tuning. According to importance of each
environment, choose different site count to do propagation model tuning, there are 4 sites
separately in Dense urban, 3 sites in urban and sub urban, and 2 sites in rural. Analyses CW test
data and output model tuning result in different environment.
The propagation models requires the definition of some general parameters, such as:

3.2
3.2.1

Frequency 874.7 MHz

Mobile antenna height : 2 m

Model for Dense Urban

Analysis CW Test Data for Dense Urban

Import the CW Test data into Atoll software, The CW signal strength of test data is following as:

Figure 3-1 CW Signal Strength plot Dense Urban

Analysis test point distribution in different clutter. Confirm test points weight in different clutter,
according to percentage, ensure which points should be reversed, which should be removed.

Site name

Number of Bin

ZTE_0219 UOB

8.883

ZTE_0215 MANDIRI

4.991

ZTE_0213 PPHUI

12.180

ZTE_0206 RESTO PAREGU

15.492

Table 3-1Test Point count of each site Dense urban

The relation between distance and signal strength

Figure 3-2 signal Strength vs 10 log (d) before filter UOB

Figure 3-3 signal Strength vs 10 log (d) before filter Mandiri

Figure 3-4 signal Strength vs 10 log (d) before filter PPHUI

Figure 3-5 signal Strength vs 10 log (d) before filter Resto Paregu

Figure 3-6 Dense Urban DT Data process

Figure 3-7 Signal strength vs log (d) after filter UOB

Figure 3-8 Signal strength vs log (d) after filter Mandiri

Figure 3-9 Signal strength vs log (d) after filter PPHUI

Figure 3-10`Signal strength vs log (d) after filter Resto Paregu

3.2.2

Model Tuning Result for Dense Urban

According to test point distribution, the relation between distance and signal and other factors,
filter out same unreasonable data points, reserved reasonable data points, to do model tuning,
output model tuning result for dense urban
The result for dense urban models are listed below:

Result
Parameter

Initial

Final

K1 (LoS)

65.4

12.13

K1(NLoS)

40

K2 (LoS)

Log(D)

65.4

42.28

K2(NLoS)

Log(D)

40

42.28

K3

Log (HTx)

-30

-19.68

K4

Diffraction

0.47

K5

Log (D) * log (HTx)

K6

K7

-5

The statistic for Dense Urban model

Statistic
Initial

Final

Mean Error

-19.82

-0.4

Standard deviation

9.72

7.78

3.3
3.3.1

Model for Urban

Analysis CW Test Data for Urban

Import the CW Test data into Atoll software, The CW signal strength of test data is following as:

Figure 3-11 CW Signal Strength plot Urban

Analysis test point distribution in different clutter. Confirm test points weight in different clutter,
according to percentage, ensure which points should be reversed, which should be removed.

Site name

Number of Bin

ZTE_0022 SALON JLO

7.801

ZTE_0083 KEMANGSELATAN

4.822

ZTE_0349 CEMENTAID

2.941

ZTE_0564 RUKOARTHAKEDOYA

5.433

Table 3-2Test Point count of each site Urban

The relation between distance and signal strength

Figure 3-12 signal Strength vs 10 log (d) before filter Salon JLO

Figure 3-13 signal Strength vs 10 log (d) before filter Kemang Selatan

Figure 3-14 signal Strength vs 10 log (d) before filter Cementaid

Figure 3-15 signal Strength vs 10 log (d) before filter Ruko Artha KEdoya

Figure 3-16 Urban DT Data process

Figure 3-17 Signal strength vs log (d) after filter Salon JLO

Figure 3-18 Signal strength vs log (d) after filter Kemang Selatan

Figure 3-19 Signal strength vs log (d) after filter Cementaid

Figure 3-20 Signal strength vs log (d) after filter Ruko Artha Kedoya

3.3.2

Model Tuning Result for Urban

According to test point distribution, the relation between distance and signal and other factors,
filter out same unreasonable data points, reserved reasonable data points, to do model tuning,
output model tuning result for Urban
The result for Urban models are listed below:

Result
Parameter

Initial

Final

K1 (LoS)

65.4

12.13

K1(NLoS)

40

28.44

K2 (LoS)

Log(D)

65.4

56.61

K2(NLoS)

Log(D)

40

48.49

K3

Log (HTx)

-30

-20

K4

Diffraction

0.3

K5

Log (D) * log (HTx)

-10

K6

K7

-5

The statistic for Urban model

Statistic
Initial

Final

Mean Error

-20.42

-0.88

Standard deviation

10.91

5.96

3.4
3.4.1

Model for Sub Urban

Analysis CW Test Data for Sub Urban

Import the CW Test data into Atoll software, The CW signal strength of test data is following as:

Figure 3-21 CW Signal Strength plot Sub Urban

Analysis test point distribution in different clutter. Confirm test points weight in different clutter,
according to percentage, ensure which points should be reversed, which should be removed.

Site name

Number of Bin

ZTE_0116 PRIMAGAMA

4.137

ZTE_3378 BINA ASIH

3.227

ZTE_4272 PERKUTUT

6.503

Table 3-3Test Point count of each site Sub Urban

The relation between distance and signal strength

Figure 3-22 signal Strength vs 10 log (d) before filter Primagama

Figure 3-23 signal Strength vs 10 log (d) before filter Bina Asih

Figure 3-24 signal Strength vs 10 log (d) before filter Perkutut

Figure 3-25 Sub Urban DT Data process

Figure 3-26 Signal strength vs log (d) after filter Primagama

Figure 3-27 Signal strength vs log (d) after filter Bina Asih

Figure 3-28 Signal strength vs log (d) after filter Perkutut

3.4.2

Model Tuning Result for Sub Urban

According to test point distribution, the relation between distance and signal and other factors,
filter out same unreasonable data points, reserved reasonable data points, to do model tuning,
output model tuning result for Sub Urban
The result for Sub Urban models are listed below:

Result
Parameter

Initial

Final

K1 (LoS)

65.4

12.13

K1(NLoS)

40

10.1

K2 (LoS)

Log(D)

65.4

34.68

K2(NLoS)

Log(D)

40

34.68

K3

Log (HTx)

-30

1.47

K4

Diffraction

0.3

K5

Log (D) * log (HTx)

-2.98

K6

K7

-5

The statistic for Sub Urban model

Statistic
Initial

Final

Mean Error

-29.83

-0.98

Standard deviation

7.15

6.19

3.5
3.5.1

Model for Rural

Analysis CW Test Data for Rural

Import the CW Test data into Atoll software, The CW signal strength of test data is following as:

Figure 3-29 CW Signal Strength plot Rural

Analysis test point distribution in different clutter. Confirm test points weight in different clutter,
according to percentage, ensure which points should be reversed, which should be removed.

Site name

Number of Bin

ZTE_2029 RANGON JAYA

8.430

ZTE_2057 SUKA TANI

4.999

Table 3-4Test Point count of each site Rural

The relation between distance and signal strength

Figure 3-30 signal Strength vs 10 log (d) before Rangon Jaya

Figure 3-31 signal Strength vs 10 log (d) before filter Sukatani

Figure 3-32 Rural DT Data process

Figure 3-33 Signal strength vs log (d) after filter Rangon Jaya

Figure 3-34 Signal strength vs log (d) after filter Sukatani

3.5.2

Model Tuning Result for Rural

According to test point distribution, the relation between distance and signal and other factors,
filter out same unreasonable data points, reserved reasonable data points, to do model tuning,
output model tuning result for Rural
The result for Rural models are listed below:

Result
Parameter

Initial

Final

K1 (LoS)

65.4

12.13

K1(NLoS)

40

12.13

K2 (LoS)

Log(D)

65.4

52.54

K2(NLoS)

Log(D)

40

52.54

K3

Log (HTx)

-30

-20

K4

Diffraction

0.28

K5

Log (D) * log (HTx)

-10

K6

K7

-5

The statistic for Rural model

Statistic
Initial

Final

Mean Error

-24.75

0.27

Standard deviation

9.91

7.83

4. LTE Propagation Model of Jakarta

The Jakarta of each model for 850 is listed in the table below
K (850 MHz)

Dense Urban

Urban

Sub Urban

Rural

K1

12.13

12.13

12.13

12.13

K2

38.63

56.61

34.68

52.54

K3

-19.68

-20

1.47

-20

K4

0.47

0.3

0.3

0.28

K5

-10

-2.93

-10

K6

K7

Table 4-0-1 Final Propagation Model tuning result of Jakarta

Appendix 1 : Equipment Description

Equipment for CW measurement are listed in table below

No

System

Equipment
CW Generator

2 Transmitter Power Amplifier

3

Omni Antena

Receiver

Receiver

5
6
7
8
9
10

Spectrum Analyzer
Site Master
Power Meter
Accesoris
Feeder Cable 7/8
Jumper
Roll Meter

SMARTFREN

Freq.

Brand

Type

850 MHz
2300 MHz
850 MHz
2300 MHz
850 MHz
2300 MHz
850 MHz
2300 MHz

HP
HP-Agilent
BVS
Minicircuit
Kathrein
FTRF
Anite Scanner
/Coyote
Agilent
Anritsu
Agilent

hp 8371 b
HP - AGILENT 8921A
BVS
ZHL 100W 242+
K736347
OA232410-NF
FSR1 /Dual Modular
Receiver
E4407B
S 331 L
Agilent

Remark

Power=40/46 dBm
Power=37/50 dBm
Gain= 11 dBi
Gain= 10 dBi

60 Meters
100 Meters

NOKIA