WO - NAST3024 - E01 - 1 UMTS Drive Test Data Analysis-72

UMTS Drive Test Data
Analysis
ZTE University
Contents
 Overview of Network Optimization Process

 Coverage-related Data Analysis
 Service-related Data Analysis
 Case Study
Engineering Optimization Phase
Ensure that network can run normally and reach preset network performance KPIs
Base station Cluster Single-site Optimization

Entire Network
Optimization Single-site optimization test
Optimization

 Cluster optimization test Site rectification
Cluster boundary optimization  Cluster optimization data analysis Single-site acceptance
Cluster optimization
and optimization plan draft
Entire network optimization and
 Cluster optimization plan
acceptance implementation
 Cluster optimization acceptance
© ZTE Corporation. All rights reserved

O&M Optimization Phase
Improve network quality and customer satisfaction
Data analysis
Network performance Optimization plan
monitoring draft
DT, CQT
OMC Complai
n
NO YES
Implementation
Test

Tool Preparation
F852 (mobile)
Supports HSDPA/UMTS
Supports EDGE/GPRS/GSM
Supports HSDPA 3.6 Mbps
downlink
Supports R99 PS 384 kbps
downlink
MF632 (Data card)

Supports HSUPA/HSDPA/UMTS

Supports EDGE/GPRS/GSM
Supports HSDPA 7.2 Mbps
downlink
Supports HSUPA 2 Mbps uplink
Supports receive diversity

Tool Preparation
 CPICH RSCP
 Coverage analysis
 CPICH Ec/Io
 Overshooting analysis
 Pilot Path Count
 Pilot pollution analysis
 PDU
 BLER
 Neighbor cell analysis
 Tx power  Call drop analysis

 Neighbours  Delay analysis
 CRC  Signaling process
 Throughput  Network KPIs
 Log
 CPICH RSCP
 Coverage analysis
 CPICH EC/Io  Overshooting analysis

 Pilot pollution analysis

Drive Test (DT)
 A drive test (DT) is to collect signal coverage data of a single site, cluster
or even the entire network as well as to collect KPIs about CS and PS
services. The data to be collected include signal strength indicator Ec,
signal quality indicator Ec/Io, service establishment success rate,
handover success rate, handover percentage, call drop rate, average
uplink/downlink throughput, delay occurred during service
establishment. All of these data helps to assess outdoor signal coverage
quality. Besides, based on these data, engineers can find out whether a
single site runs normally and locate faults that cause network KPIs
deteriorate, such as device faults, pilot pollution, cell faults, improper
settings of system parameters, and so on.

Call Quality Test (CQT)
 A call quality test (CQT) is a test towards a fixed point. Its
purpose is to collect indoor and outdoor signal strength data as
well as to collect data about service establishment and
maintaining.
 A fixed point test is to check signal quality of a local area. KPIs of
service establishment and maintaining are useful for analyzing
network problems occur at the fixed point, especially problems
like indoor/outdoor signal optimization, repeater interference,
complexity of radio propagation environment, and so on.
 Key KPIs that the test focuses include link signal quality,
multipath impact, service establishment status, cell throughput,
and so on. compared to DT, it does not care about KPIs about
handover.

Contents

 Case Study
Pilot Pollution
 Cell pilot signals received at a test point have no difference in
signal strength (either strong or weak), but no dominant pilot is
received.
 The symptom is that the received pilot power is good but Ec/Io
of each cell is weak.
 At present, most WCDMA equiment supports up to three active
sets. If the number of similar Ec/Io values of different cells is
more than three, it can be regarded as a kind interference
towards three radio links in the active sets.
 The causes of signal distortion include, overshooting of sites at
higher places, ring deployment, streect effect, strong reflectors,
and so on.
 Pilot pollution algorithm
 Algorithm1: Ec/Io > Ec/Io_for_BestServingCell - the number of 8dB
cells is more than 3
 Algorithm2: Ec/Io > Ec/Io> Ec/Io_for_BestServingCell - the number
of Window_add cells is more than Active Size

Pilot Pollution
Check for pilot pollution via the test data of Scanner
 Pilots in Scanner, rscp>Trscp & ecio > Tecio

Pilot Pollution
Check for pilot pollution via the test data of UE-method 1

Pilot Pollution
Check for pilot pollution via the test data of UE-method 2

Neighbor Cells without Configuration
 A neighbor cell may has been added to an active set. It may be
involved in cell re-selection or handover with other cells.
 When a mobile is in the idle and FACH state, the mobile reads
the neighbor cell information from system messages, and
measures these neighbor cells when conditions are met. If a
neighbor cell fails to be set, cell re-selection cannot be complete
timely which further cause call establishment failure.
 When the mobile is in the DCH state, RNC sends out neighbor
cell information through measurement control command to ask
the mobile to measure all neighbor cells where handover may
occur. In this case, if a neighbor cell fails to be set, handover
failures may occur and further cause call drops.

Compare Ec/Io results obtained by the mobile and that obtained by Scanner to see
whether a cell is set or not
 Ec/Io results obtained by UE are poor while that obtained by Scanner are good.
 Scanner recorded that Best Server scramble did not appear in the Measurement
control signaling sent out by RNC.

Judge whether a cell is set or not via the active set and monitor set measured by the mobile
 There are cells whose signal is strong in the monitor set,

but they are not added to the active set.
Unreasonable Handover Area
 Typical handover process: measure control － > measure report
－ > handover dicision － > handover implementation － > new
measurement control.
 Unreasonable handover area may cause that signals jitter to fast
and have not enough time to complete handover which further
causes call drops.
 Corners and street crossroads

Unreasonable Handover Area
Compare Ec/Io results obtained by the mobile and that obtained by Scanner to
see whether a cell is set or not
 Ec/Io results obtained by UE are poor while that obtained by Scanner are good.
 Scanner recorded that Best Server scramble did appear in the Measurement
control signaling.
Pilot Pollution
Solutions
 The core to solve pilot pollution is to form dominant pilot at the

place where pilot pollution occurs.
 Optimization methods include:
 Adjust antenna engineering parameters, like azimuth, downtilt
angle, installation position (install antenna at a higher place)
 Adjust cell pilot transmission power, including increasing
transmission power of a certain cell and decreasing that of other
cells
 Adjust base station layout, and add signal sources in pilot
pollution areas to introduce a strong primary signal
 Adjust cell selection and re-selection parameters when necessary
to imrpove call access success rate

Overshooting
 Pilot signals in some area are too strong and cover the areas beyond their planned coverage range.
This leads to the formation of discontinuous dominant areas in the coverage range of other cells.
 The places where bases stations are deployed are too high; antenna engineering parameters are not
reasonable; street effects; and so on.
Use Best SC to check

for overshooting

Overshooting
Use Over Shoot to check for
overshooting

Overshooting
Use Over Shoot to check for overshooting
 Global sample points: RSCP > -105dBm & Ec/Io >-15dB

 Overshooting sample points:
 Distance from a sample point to a cell > Geographical threshold (Distance)
 RSCP > RSCP.Threshold & Ec/Io > Ec/Io.Threshold
 If number of global sample points >M & number of overshooting
sample points >N & number of global sample points / number of
overshooting sample points > Propotion, then it indicates that there are
overshooting problems.
 Measure of severity
 Measure of distance: W1=distance from a sample point to a cell /
geographical threshold
 Measure of severity:

W2=W21, indicates that ecio of an overshooting sample point is strongest

W2=W22, indicates that ecio of an overshooting sample point is the second
strongest

W2=W23, others
 Overshooting degree: W=W1*W2

Contents

 Case Study
RRC Connection Establishment Success Rate
RRC connection establishment success rate (service-related)
＝ Number of RRC connection establishment success times (service-related)

/Number of RRC connection establishment attempts (service-related)×100%

RRC Connection Establishment Success Rate
RRC connection
established
NAS connection
established
Authentication
Encryption
Call established
RB established
Conversation
Release of the called party
RRC connection released

Call Access Success Rate
Call access establishment success rate

＝ Number of RB establishment success times /
Number of RRC connection establishment attempt times × 100%

Call access establishment success rate
RRC connection established
NAS connection established
Authentication
Encryption
Call established
RB established
Conversation

Call Establishment Success Rate
Call establishment success rate

＝ Number of call establishment success times /
Number of RRC connection establishment attempt times × 100%

Call establishment success rate

Authentication
Encryption
Call established
RB established
Conversation

PDP Context Activation Success Rate
PDP context activation success rate

＝ Number of PDP context activation success times /
Number of PDP context activation attempt times
×100%

PDP Context Activation Success Rate
RRC connection
established
NAS connection
established
Authentication
Encryption
Call established
RB established
Conversation
PDP deactivated

Average Access Delay
Average access delay

＝ { Time that UE receives RRC Connection Setup-
Time UE sends RRC Connection Request}/
Number of RRC connection establishment attempt times

Average access delay
RRC connection
established
NAS connection
established
Authentication
Encryption
Call established
RB established
Conversation

Channel Synchronization Delay
Channel synchronization delay
＝ { Time UE sends RRC Connection Setup Complete-

Time UE sends RRC Connection Request}/
Number of RRC connection establishment attempt times

Channel synchronization delay

Authentication
Encryption
Call established
RB established
Conversation

Call Access Delay
Call access delay
＝ { Time the calling UE sends Radio Bearer Setup Complete-

Time the calling UE sends RRC Connection Request}/
Number of call attempt times

Call access delay

Authentication
Encryption
Call established
RB established
Conversation

Call Establishment Delay
Call Establishment delay
＝ { Time calling UE receives Alerting-

Time the calling UE sends RRC Connection Request}/
Number of call attempt times

Call Establishment Delay

Authentication
Encryption
Call established
RB established
Conversation

PDP Context Activation Delay
PDP context activation delay
＝ { Time UEA receives Active PDP Context Accept-

Time calling UE sends Active PDP Context Request}/
Number of PDP activation attempt times

PDP Context Activation Delay
Authentication
Encryption
Call established
RB established
Conversation
PDP deactivated

Contents

 Case Study
Locating Area Where Call Drop Occurs
Use Marker Setting to locate the area
Select calling events like RRC Connect Fail, Call Drop, RB setup Failure.
Then the system displays the area where call drop occurs in the Map window.

Check if the signaling is complete to locate the area
Select calling events like RRC Connect Fail, Call Drop, RB setup Failure.
Then the system displays the area where call drop occurs in the Map window.

Check if the signaling is complete to locate the area
Broadcasting messages are received suddenly; call re-establishment fails.

Check if the signaling is
complete to locate the area
 Radio quality
deteriorates;
 UE initiates cell update
process and waits for
call re-establishment
 The network re-
configure the
transmission channel
for the UE through
Cellupdate Confirm to
re-synchronize radio
links;
 Call re-established
successfully.
Broadcasting messages are received suddenly; call re-establishment

succeeded without call drops.
Poor RSCP

Before Optimization
The signal is blocked by buildings.

Optimization Plan
Analyze the site survey photos, and then

Change antenna direction of SC283 from 160 to 140 degree.
Change antenna downtilt of SC283 from 7 to 2 degree.

After Optimization
Optimization result: coverage is improved to some extent.

Poor Ec/Io

Before Optimization
505
Problem: some Ec/Io in the red circle is poor below -10dB, RSCP is normal around -
85dBm
Optimization Plan
 The red circle area, which is in the middle of several NodeBs, has
no primary server cell for coverage. In order to improve Ec/Io, it
is necessary to select a cell, SC505 in the case, as primary server
cell.
 Change antenna direction of SC505 from 130 to 100 degree.

 Change antenna downtilt of SC505 from 6 to 4 degree.

After Optimization
Optimization result: The primary server cell is more outstanding, and Ec/Io is improved
obviously.

Problem: low HSDPA download rate

Before Optimization
13
20
18
Problem: The red circle area, active set has three cells Ec/Io all below -10dB, RSCP is
around -85dBm, causing HSDPA download rate lower than 1Mbps.

Optimization Plan
 The main idea is to make primary server cell outstanding. After

analysis on the basis of field environment, consider SC18 to cover
the red circle area. And decrease the effect of SC20 and SC13,
making them not cover the area.
 Change antenna direction of SC18 from 90 to 75 degree.
 Change antenna downtilt of SC20 two degree more.

Pilot Pollution
 Results to network:
 decreasing the CSSR (Call Setup Success Rate)
 decreasing the SHO rate
 decreasing the system capacity
 Trouble shooting
 Generating a dominating cell in the pilot pollution area

After Optimization
18
Optimization result: primary server cell S18 is outstanding. Ec/Io is improved obviously
and HSDPA download rate increase to normal above 2Mbps.

Optimization Issues
 The hill blocking signal causes call drop

 Overshooting at Sofitel Beach
 No Throughput Round SWHT Site for R99 Ftp Download

Issue1: The hill block signal cause call drop
site cell signal are blocked by

a mountain, UE lock in 3G
cause call drop in the back of
mountain.

Before Optimization
Problem: The signal of this is area quite poor because the mountain blocked signal
from site, SC27 cell RSCP<-95dBm, if UE lock in 3G and can not handover to 2G, it
will cause call drop in the back of the mountain.

Optimization Plan
 Change 2d event threshold (compress mode) of SC27 cell from -

95dBm to -92dBm.
 Change 3a event threshold (inter-RAT handover) of SC27 from -
95dBm to -92dBm.
 Change 3a time to trigger of SC27 from 100ms to 80ms.
 Purpose: Let UE handover to 2G more faster in order to avoid call
drop.

After Optimization
Optimization result: UE can success to handover from 3G to 2G if mobile choices “auto

mode”, the result show no call drop.

Issue2:
Overshooting
Signals of HNSS2 and

KHTK1 were reflected by
sea and result in
overshooting at Sofitel
Beach.

Before Optimization
17
20
11
Problem: There are overshooting signal of SC11(KHTK1) and SC17(HNSS2) in the

Sofitel Beach. Distances to sites are over warn distance, customer can not share 3G
service, including CS and PS.

Optimization Plan
 Change antenna direction of SC17(HNSS2) from 180 to 200

degree.
 Change antenna E-downtilt of SC17(HNSS2) from 2 to 5 degree.
 Change antenna direction of SC20(SWHT2) from 190 to 170
degree.
 Change antenna E-downtilt of SC11(KHTK1) from 6 to 8 degree.
 Add neighbor cells among SC17(HNSS2), SC11(KHTK1) and
SC20(SWHT2).

After Optimization
20
Optimization result: SC20(SWHT2) is primary service cell at mostly area in the beach.
Though signals of SC 17(SNSS2) and SC11(KHTK1) still exist in the beach, all service (CS
and PS) are normal.

Issue3:No throughput Round SWHT Site for R99
Ftp Download
Soft handover failure cause

low throughput for R99
service.

Before Optimization
Problem: Signal quality of SC28(SWHT3) is better than signal quality of SC12(SWHT1), but
UE can not report 1A event, SC28 cell can not update to active set , so result in R99
throughput is low.
Optimization plan: First step check whether NodeB has alarm , the second check RNC
parameter about intra and find that intra parameters set isn’t correct set, so change back
to correct parameters set and resolve problem.
After Optimization
Optimization result: SC28(SWHT3) can be added into active set, soft handover is
success and R99 throughput is normal.

WO - NAST3024 - E01 - 1 UMTS Drive Test Data Analysis-72

Uploaded by

Copyright:

Available Formats

WO - NAST3024 - E01 - 1 UMTS Drive Test Data Analysis-72

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

WO - NAST3024 - E01 - 1 UMTS Drive Test Data Analysis-72

Uploaded by

Copyright:

Available Formats

UMTS Drive Test Data

 Overview of Network Optimization Process

Base station Cluster Single-site Optimization

© ZTE Corporation. All rights reserved

© ZTE Corporation. All rights reserved

Supports HSDPA 3.6 Mbps

MF632 (Data card)

Supports HSDPA 7.2 Mbps

Supports receive diversity

© ZTE Corporation. All rights reserved

 Tx power  Call drop analysis

 CPICH EC/Io  Overshooting analysis

© ZTE Corporation. All rights reserved

© ZTE Corporation. All rights reserved

© ZTE Corporation. All rights reserved

 Overview of Network Optimization Process

© ZTE Corporation. All rights reserved

 Pilots in Scanner, rscp>Trscp & ecio > Tecio

© ZTE Corporation. All rights reserved

© ZTE Corporation. All rights reserved

© ZTE Corporation. All rights reserved

© ZTE Corporation. All rights reserved

© ZTE Corporation. All rights reserved

 There are cells whose signal is strong in the monitor set,

© ZTE Corporation. All rights reserved

 The core to solve pilot pollution is to form dominant pilot at the

© ZTE Corporation. All rights reserved

Use Best SC to check

© ZTE Corporation. All rights reserved

© ZTE Corporation. All rights reserved

 Global sample points: RSCP > -105dBm & Ec/Io >-15dB

© ZTE Corporation. All rights reserved

 Overview of Network Optimization Process

RRC connection establishment success rate (service-related)

＝ Number of RRC connection establishment success times (service-related)

© ZTE Corporation. All rights reserved

Release of the called party

RRC connection released

© ZTE Corporation. All rights reserved

Call access establishment success rate

© ZTE Corporation. All rights reserved

RRC connection established

NAS connection established

Release of the called party

RRC connection released

© ZTE Corporation. All rights reserved

Call establishment success rate

© ZTE Corporation. All rights reserved

RRC connection established

NAS connection established

Release of the called party

RRC connection released

© ZTE Corporation. All rights reserved

PDP context activation success rate

© ZTE Corporation. All rights reserved

RRC connection released

© ZTE Corporation. All rights reserved

Average access delay

© ZTE Corporation. All rights reserved