6/29/2015

Agenda
Comparison between LTE
and CDMA

How to optimize the

planning

Cell Coverage Theory Analysis

Different Elements Same Elements

Technologies Content TDD-LTE CDMA-Data Technologies Content TDD-LTE CDMA-Data

1 data rate 512 153.6 1 Antenna Height 35 35

NodeB Antenna
2 SINR/EbNo -4.5 0.49 2 18 18
Gain
NodeB Sensitivity
3 -108 -118.00
（dBm）
4 Frequency 2600 800
K1 158 K1 135
5 Propagation Model
K2 40.9 K2 44.9
6 UE CPE Telephone

7 Penetration 15 12

8 UE Height 3 1.5

9 Handover gain 2.76 3.26

10 Path loss 131 136

Why the coverage range of TDDLTE is less than CDMA?

1
 6/29/2015

Radius Result-Suburban
Uplink Radius C800MHz EVDO
153.6Kbps
（ km）
Radius（ ） C800MHz 语音
7.00

8.00 6.00 G900MHz 语音

6.00
5.00
4.00 FDD2.1GHz 512Kbps
2.00 4.00
0.00
3.00 TDD1.9GHz 512Kbps
2.00
TDD2.3GHz 512Kbps
1.00

0.00 TDD2.6GHz 512Kbps

SU

CDMA800M-Data
Technologies TDDLTE-2.6G CDMA800M-1X Voice
Service
Radius（km） 1.14 4.92 7.00
Ratio 1 4.3 6

Site Scale
R=1.14 R=4.92
D

LTE S(Area)=1.95*R*R CDMA

Sites
CDMA800M-Data
12% Technologies TDDLTE-2.6G
Service
LTE Sites Radius（km） 1.14 4.92
88% CDMA Sites Sites 53 7

2
 6/29/2015

Agenda
Comparison between LTE and
CDMA

How to optimize the planning

Assumptions for simulation

1、When we do the
simulations, it’s hard to
assume effective offset of
each different clutter types
other wise we can collect the
real DT test data in
implementation stages.

2、Each sector and each site

will have more than 1 clutter
types, previously we assume
one clutter type for one site
or more similar sites, now
we’d like to make the
improvement for K1-K7 and
clutters’ offset.

3
 6/29/2015

Propagation Model Values and forum for simulation tools

For simulation part, because all of the points only follow one
propagation forum.
Ploss = K 1 + K 2 log( d km ) + K 3 H meff + K 4 log( H meff ) + K 5 log( H eff ) +
K 6 log( H eff ) log( d km ) + K 7 Diffn + K clutters
RSRP =RS signal Tx power + BS antenna gain- Ploss-indoor loss- body loss-
cable loss- shadow margin+ UE antenna gain

1、most of sites use the same

Propagation Model.
2、All the simulation pixel use
the same Mathematic Forum.
3、So the RSRP indictor from
simulation are almost the
same trend for each site and
direction.

Model Mathematic
Limit Forum limit

DT Test guide Planning to tune the Propagation Model

DT test guide planning
to fine tune the
propagation model

Give suggestions to
DT test, such as
azimuth and down tilt.

Find different suitable K

values for different clutter
types.

With the project going on, we’d like to optimize our planning
parameters. This work will also help us in the rest of the
projects.

4
 6/29/2015

Improvement Steps of Planning Workflow

Network Workflow
Requirement Nominal Planning Stage
Analysis(input 1
Requirement data)

Coverage and Detailed Planning Stage

Dimensioning Capacity 2
Dimensioning

simulation Network 3
simulation
Model Tuning
DT Test Data
Parameters 5
Site survey Optimization 4
Suitable Model for each
clutter
Optimization DT test

Useful resource data for next

step of planning work
9

Parameters Optimization

5
 6/29/2015

Pre-Simulation,RSRP Post-Simulation，RSRP

Average:-
103dBm@95%

Propagation Model Tuning

6
 6/29/2015

Procedure of Propagation Model Tuning

Different country will have different
morphology
1、Define Clutter
Types

2、Define Test route

3、collect Test data

4、Import data to CXP

tools

5、Analyze data and

filter the unnecessary
sample

6、Analyze and get

the new value

Clutter Type Classification

Urban Dense • High building Area

Urban Vegetation • Trees and forest in urban area

Urban Hilly • More residential around hilly

Flat Open Region • Open area

Suburban Vegetation • Small village

Suburban Hilly • Few people around hilly

Lake City • Lake or river area

7
 6/29/2015

Theory of Propagation Model Tuning

1、First, we collect the DT test data and filter the data resource.
2、Second, we find the most accurate K value for all the simulation pixel.
3、Third, during the tuning, we make sure the Standard Deviation is lower than 10dB.
The whole match degree is reflected by the standard deviation value.
Yi=X1i-X2i X1i perdition, X2i test

Mean Error(u)=

Standard
Deviation

0 1 2 3 4 5 6 7 8 9 10 11 12
Standard Deviation International standard
Ploss = K 1update + K 2 update log( d km ) + K 3 H meff + K 4 log( H meff ) + K 5 log( H eff ) +
K 6 log( H eff ) log( d km ) + K 7 Diffn + K clutters − update

Site B-AW_Awissawella

It’s a suburban site, there are dense

vegetation around it.
Now, after the model tuning , the mean
error is 0dB, and Standard Deviation is
6.1dB.

8
 6/29/2015

Simulation and DT Test

DT Test

Matched Area

Clutter offset

9
 6/29/2015

Site B-NG_Badalgama

Site B has the similar

morphologies compared
with Site A, so after we
optimize the Model for Site
A, then use it for Site B, we
find that the model also fit
to Site B.

Thank you!

10