Rf Optimisation

ICN PLM CA NP

© SIEMENS Limited 1999

s

Main Topics What is network optimisation? Why optimisation? Aim of network optimisation Advantages for the customer Planning vs. optimising Major problem areas Radio optimisation related processes Tuning Test types Measurement analysis Change request and action Acceptance tests Ongoing optimising Pre-analysis: general network check Customer complaints analysis Collect/analyse OMC statistics Collect/analyse drive test measurements Implement changes Test mobile Repeated call setups Continuous call Statistics

ICN PLM CA NP

2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 10 11 11 12 13 14

Concept for optimisation Analysis programs Problem symptoms Coverage analysis Test mobile measurements Possible problem areas Antenna configuration Antenna types - typical beam patterns Antenna fine tuning Omni vs. sectorised Vertical antenna beam Tilting Antennadiversity type Verification of RF network design Site check Antenna isolation Site physical configuration Site-to-site distances and distribution Special features for improving coverage Cell splitting, sectorisation DTM check Propagation model verification Link budget analysis

© SIEMENS Limited 1999

14 15 15 16 16 17 18 18 19 20 20 21 22 23 23 24 25 25 26 26 27 27 28

s

Main Topics (continued) Dropped call analysis Call setup analysis MAXRETR Handover performance analysis Handover parameters Consequence of missing neighbours Consequence of many neighbour definitions Handover measurements Handover parameters Radio link measurements Handover algorithm Handover criteria - quality Handover decision Intracell handover Level handovers Distance handover Power budget handover Cell reselection Speech quality analysis Downlink interference measurement Frequency changes BSIC optimisation

ICN PLM CA NP

29 30 30 31 31 32 32 33 33 34 36 37 37 38 39 40 40 41 43 43 44 45

Call setup/handover mechanisms Location area codes Interference reduction Power control Frequency hopping DTX Channel configuration Capacity enhancements Adding TRX Interference reduction features Traffic load distribution Call setup/handover mechanisms Hierarchical cell structures Concentric cells Overlaid micro-and picocells Microcell frequency planning Speed sensitive handovers Half rate coding/dual rate operation Cell parameter optimisation Performance measurements

© SIEMENS Limited 1999

45 46 46 47 48 49 50 50 51 51 52 52 53 53 54 54 55 55 56 56

s

What is Network Optimisation? Improving Capacity, Quality and General Performance of the existing Network Infrastructure

ICN PLM CA NP

© SIEMENS Limited 1999

s

Why Optimisation? Coverage holes Performance degradation by interference Different subscriber distribution compared to that assumed for the network design Unexpectedly high subscriber growth Extensive network expansions ongoing Frequency resources at the limit Unexpected mobility profile of subscribers

ICN PLM CA NP

© SIEMENS Limited 1999

s

Aim of Network Optimisation Improved Network Quality ➨

Speech quality, Call success rate, Call setup time

Improved Network Availability ➨

Service area , Radio Coverage

Optimised utilisation of installed equipment ➨

Increase in subscriber potential

ICN PLM CA NP

© SIEMENS Limited 1999

s

Advantages for the Customer Optimum utilization of the system resources Minimized costs

Reduced subscriber complaints Optimised subscriber satisfaction

Increased Profit

ICN PLM CA NP

One step ahead of the Competitors

© SIEMENS Limited 1999

s

Planning vs. Optimising Thorough network planning from start can reduce the optimisation effort significantly! ➨ In a poorly planned network, achievable optimisation effects without major re-design are rather marginal A close link between the two activities is necessary Be involved

Feedback result

ICN PLM CA NP

© SIEMENS Limited 1999

s

Major Problem Areas no coverage interference blocking handover not working HW/SW failures

ICN PLM CA NP

© SIEMENS Limited 1999

s

Radio Optimisation Related Processes Tuning

Acceptance Tests

Ongoing Optimisation

The following processes involve optimisation related activities ➨

Tuning Process ✲ ✲

➨ ➨

drive tests adjustment of network parameters

Acceptance tests Ongoing Optimisation ✲

ICN PLM CA NP

Repeated quality control and improvement as network grows / matures

© SIEMENS Limited 1999

s

Tuning Test Measurement

Measurement Analyzing

Change Request Action

Repeat Process until Agreed Quality

Objectives : ➨ ➨

➨

ICN PLM CA NP

Verify network configuration against current planning status Identify and eliminate equipment faults (HW/SW) and installation errors Ensure that the network is ready for acceptance testing

© SIEMENS Limited 1999

s

Test Types Continuous drive test ➨

setup a test call and drive over an area for detecting lack of coverage, missing handovers, interferences etc.

Spot test ➨

detail measurement to be taken at dedicated problem spots for detail analyzing of specific problem

ICN PLM CA NP

© SIEMENS Limited 1999

s

Measurement Analysis Antenna Installation check ➨

height, orientation and tilt

Basic cell parameters and functions ➨

OMC ✲ ✲ ✲ ✲ ✲

BCCH, BSIC, CI, LAC Neighbour List, consistency HO and power parameters Call Setup on all timeslots and speech quality check HO to other sectors or other neighbours

Test measurement (TEMS etc. together with a GPS) ➨ ➨ ➨ ICN PLM CA NP

Signal Strength Co-channel and adjacent interference Handover relations © SIEMENS Limited 1999

s

Change Request and Action SBS System Database ➨ ➨ ➨

Change BCCH to avoid interference Change HO-Margin Add neighbour relations (Mutual)

Site Hardware ➨

Antenna tilt etc.

System error ➨ ➨

Software bugs Transmission sync. (ADPCM)

ICN PLM CA NP

© SIEMENS Limited 1999

s

Acceptance Tests Setup Test Scenario

Performing Test

Setup Test Scenario ➨ ➨

✲ ✲ ➨ ➨ ➨

ICN PLM CA NP

✲ ✲

Test Purpose Test Definitions ✲

Coverage Criteria Coverage Area Successful Call

Test Result

✲ ➨

Test Analysis ✲

➨

✲ ✲ ✲

© SIEMENS Limited 1999

Acceptance Criteria

Test Results ✲

Test Condition Test Equipment Test Methodology

Test Routes Test Procedure Test Duration

Signal Level Signal Quality Handover Call Success Rate

s

Ongoing Optimising For improvement of the network after it is launched and filled up by subscribers Pre-analysis: Pre-analysis: General General network network check check

Collect Collect/ / analyse analyse complaints complaints

Collect Collect/ / analyse analyse OMC OMC statistics statistics

Repeat Repeat process processuntil until agreed quality agreed quality

ICN PLM CA NP

© SIEMENS Limited 1999

Collect Collect/ / analyse analyse drive drivetest test measuremts measuremts

Propose Propose/ / implement implement changes changes

s

Pre-analysis: General Network Check Steps to be carried out: ➨ ➨ ➨ ➨

Kick-off meeting Determine original network planning objectives Collect information about network status Determine functional network structure, e.g. – - BTS / BSC locations., antenna direction etc. – - services and features used – - network structure (macrocell, microcell etc.)

➨

Determine the network element configuration, e.g. – - number of TRX per cell – - sector / omni config.

➨ ➨ ICN PLM CA NP

Visit selected sites (if necessary) Database analysis © SIEMENS Limited 1999

s

Customer Complaints Analysis Additional source of information, but difficult to handle Customer service desk must collect all relevant information ➨ ➨ ➨ ➨ ➨

Caller and Called No. (PSTN->MS, etc.) What is the problem? (Voice Quality, Can’t make a call, etc.) MS is moving or fixed while make call Where did the problem occur? When?

ICN PLM CA NP

© SIEMENS Limited 1999

s

Collect / Analyse OMC Statistics OMC Measurement ➨ ➨ ➨ ➨

Handled traffic (congestion on TCH, SDCCH) dropped calls Interference Handover reason (due to UL_QUAL, Powerbudget, distance…)

Advantages over test drives: ➨ ➨ ➨ ➨ ➨

Less labor intensive and time consuming More comprehensive, based on large number of users not limited to time of test drive Uplink and Downlink analysis possible Subscriber behavior mix of outdoor, indoor, incar use

ICN PLM CA NP

© SIEMENS Limited 1999

s

Collect / Analyse OMC Statistics Disadvantages, limitations: ➨ ➨

Limited geographical resolution (Where does the problem occur?) Cannot separate problems due to coverage from other ✲ ✲

➨

Call attempts in uncovered areas are not counted Call drop due to lack of coverage

Network must have minimum load for reliable statistics

ICN PLM CA NP

© SIEMENS Limited 1999

s

Collect / Analyse Drive Test Measurements Test types ➨ ➨ ➨

Continuos drive test (Trace mode) Spot test Network performance test (Statistical mode)

Test Measurement ➨

Collect MS measurement report data (Downlink only!!) ✲ ✲ ✲ ✲

ICN PLM CA NP

Serving signal level BER (Rxqual) Channel Number CI and LAI

✲ ✲ ✲ ✲

Timing Advance Layer 3 messages BSICs Signal and power levels of neighbouring cells

© SIEMENS Limited 1999

s

Implement Changes Changes related to database parameters Actions related to site hardware Problems to be solved by Normal Roll-out activities Problems to be solved by other system experts

ICN PLM CA NP

© SIEMENS Limited 1999

s

Test Mobile Various modes, e.g. ➨ ➨ ➨

Repeated call setups Continuous call Scanning mode ✲

check for spectrum

occupancy ✲

ICN PLM CA NP

check for BCCH with no neighbour relations

© SIEMENS Limited 1999

s

Repeated Call Setups Um-

Abis-

Base Transceiver Station

Method ➨ ➨

Base Station Controller

✲

Measuremt Software

predefined time = mean holding time call may be dropped earlier

repeat call setup after predefined waiting time (typical 15 s)

Purpose ➨ ➨

Mobile Switching Center

call setup hold for predefined time period and then release ✲

➨

Serial

A-Interface

simulate subscriber behavior wide area quality assessment and trend identification

ICN PLM CA NP

© SIEMENS Limited 1999

PSTNInterface

s

Repeated Call Setups Typical parameters ➨ ➨

call setup success rate, setup time, dropped call rate statistics can be generated in Tornado / Planet, e.g. Call Diagnostics RxQual Full Threshold: RxQual Full Threshold (%): 90 RxLev Full Threshold: RxLev Full Threshold (%): 90 Maximum Setup Time (s): Call 1 2 3 4 5 6

ICN PLM CA NP

Time 21:38.8 23:53.1 26:08.7 28:23.9 30:38.8 32:54.4

Setup 6.5 FAIL 5.7 6.4 5.8 12

4 14 10

Clear Down RxQual (%) RxLev (%) Category OK 100 100 GOOD FAIL FAIL FAIL NO SETUP OK 98 85.3 LOW SIGNAL OK 79.5 100 NOISY FAIL FAIL FAIL DROPPED OK 100 100 DELAYED

© SIEMENS Limited 1999

s

Continuous Call Method ➨ ➨

call setup hold continuously until drive test route complete ✲

in case of call drops re-establish

Purpose ➨ ➨ ➨ ➨ ➨

Wide area quality trace Locating individual problem areas Detailed analysis in problem areas Quality assessment on rural highways etc. BS Testing and Functional Testing

ICN PLM CA NP

© SIEMENS Limited 1999

s

Continuous Call Typical parameters ➨

RxLev, RxQual, BCCH, BSIC, handover, Layer 3 messages etc.

Import into planning tool ➨ ➨

Terrain or clutter background Comparison of measured network performance vs. prediction

Statistics: ➨

RxLev, RxQual, handover success rate

ICN PLM CA NP

© SIEMENS Limited 1999

s

Statistics Combine from both modes Measurement RxLev > -85 dBm RxQual < 4 Handover success rate Call setup success rate Mean setup time Dropped call rate

Test sample unit No. of samples Measured value Measurement bin (Tornado) 8,432 99.90% Measurement bin (Tornado) 8,432 99.20% Handover attempt 61 93.50% Call attempt 115 90.30% Call successfully setup 106 5.3 s Call successfully setup 106 1.00%

Typical measurements also used for acceptance tests

ICN PLM CA NP

© SIEMENS Limited 1999

s

Performance Measurements Provide an overview of network performance (statistics) ➨ ➨

uplink analysis also possible validity depends on sufficient samples

Examples: ➨

blocking rate BTS ID LAC 6 2 5 22 1

4 4 4 4 4

BTS ID LAC 25 6 26 3 ICN PLM CA NP

4 4 4 4

CI

BSIC

4052 4083 4051 4183 4082

2 2 2 2 2

CI

BSIC

4052 4052 4171 4041

2 2 2 2

4 2 6 0 1

4 4 6 7

f1

f2

83 76 79 77 84

69 67 66

f1

f2

f3

f4

80 f3

f4

83 69 83 69 63 87 © SIEMENS Limited 1999

Busy hour

TCH Blocking Rate

16:00:00 16:00:00 16:00:00 12:00:00 13:00:00

66.53% 30.16% 7.91% 3.96% 3.81%

Busy hour SDCCH Blocking Rate 15:00:00 16:00:00 13:00:00 13:00:00

32.99% 5.99% 2.83% 2.06%

s

Performance Measurements ➨

➨

Call setup success rate BTS ID

LAC

CI

Busy Hour

Call Set-up Success Rate

25 29 15 5 26 11

4 4 4 4 4 4

4152 4131 4032 4051 4171 4071

15:00:00 15:00:00 18:00:00 16:00:00 13:00:00 12:00:00

28.4% 68.0% 81.3% 92.1% 94.1% 94.7%

Dropped call rate BTS ID LAC 37 15 22 25 7 26 29 27 19

ICN PLM CA NP

4 4 4 4 4 4 4 4 4

CI 4192 4032 4183 4152 4011 4171 4131 4172 4212

TCH RF Loss Inter Cell HO Connections Loss 19730 12740 10993 24748 8849 15922 5712 10421 9192

1526 723 485 755 240 219 77 156 130

23 6 18 12 16 28 8 4 9

© SIEMENS Limited 1999

Intra Cell HO Loss

Call Drop Rate

153 58 13 29 23 12 6 4 5

9% 6% 5% 3% 3% 2% 2% 2% 2%

s

Concept for Optimisation

pr im to w rk? Ho two ne

Alternatives • Status of the Network • Decide further Analysis Program

Analyzing Programs Coverage Dropped Calls

e ov

Network Snapshot

the

Call Setup Success Handover Perf. Speech Quality

Quick Check ICN PLM CA NP

General Check

© SIEMENS Limited 1999

s

Analysis Programs Coverage:

Analysis for Fulfilment of Coverage Requirements (Urban, rural ... areas, outdoor, in-car, indoor)

Dropped Call:

Analysis for Dropped Calls due to Interference, SW/HW failures, Transmission Network Failures

Call Setup:

Analysis for Blocking and Capacity Limitations, Analysis for Resource Allocation Procedures

Handover:

Analysis for Efficient Handover Performance

Speech Quality:

Analysis for Interference

ICN PLM CA NP

© SIEMENS Limited 1999

s

Problem Symptoms No service

High call drop rate No coverage RF Network No System Availability No coverage Network Element Failures Interference Transmission Network Failures Handover failure Fixed Network BSS, SSS Network Element Failure Low call setup success rate Transmission Failures RF Network Other networks No coverage Mobile terminal Interference Blocking Poor speech quality Fixed Network BSS, SSS RF Network Blocking No coverage Overload Interference Other Poor handover performance Fixed Network BSS, SSS Network element failure Transmission network failure Other networks Mobile Phone ICN PLM CA NP

© SIEMENS Limited 1999

s

Coverage Analysis Test mobile measurements Antenna configuration check Verification of RF network design DTM check Propagation model verification Link budget analysis

ICN PLM CA NP

© SIEMENS Limited 1999

s

Test Mobile Measurements Collect RxLev measurements together with GPS co-ordinates Analyse on planning tool Reasons for poor coverage: ➨

serving cell not best server ✲

➨

handover problems

best server signal low ✲

check site / network design

Analyse in terms of relevant thresholds: ➨ ➨ ➨

indoor level in-car level outdoor level

ICN PLM CA NP

© SIEMENS Limited 1999

s

Test Mobile Measurements Consequences of poor RxLev: ➨ low RxQual ➨ vulnerable to interference Limitation with drive tests: ➨ downlink only Another method: ➨ statistical analysis ➨ OMC or drive tests ICN PLM CA NP

© SIEMENS Limited 1999

s

Possible Problem Areas Downlink ➨ ➨ ➨ ➨

➨

Uplink

Output power low Obstruction of Tx antenna Antennae not aligned properly Broken / wrongly connected cables Database parameters controlling output power

ICN PLM CA NP

➨

➨ ➨ ➨

➨

© SIEMENS Limited 1999

Receive sensitivity degraded due to hardware problems Obstruction of Rx antennae Antennae not aligned properly Broken / wrongly connected cables Lack of diversity gain

s

Antenna Configuration General points to check ➨

antenna type, e.g. ✲ ✲ ✲ ✲

➨

antenna azimuth angle (for directional antennae) ✲

➨

coverage targets

antenna tilt angle ✲

➨

omni directional 60, 90 or 120 degrees electrical downtilt cross-polarised

electrical + mechanical

diversity & isolation ✲ ✲

ICN PLM CA NP

e.g. space diversity, polarisation diversity © SIEMENS Limited 1999

s

Antenna Types - Typical Beam Patterns Directional antenna

ICN PLM CA NP

© SIEMENS Limited 1999

s

Antenna Types - Typical Beam Patterns Omni antenna with electrical downtilt

ICN PLM CA NP

© SIEMENS Limited 1999

s

Antenna Fine Tuning Horizontal Plane: ➨ ➨ ➨

Possible coverage weakness between sectors Interference reduction Traffic load distribution

Vertical Plane: ➨ ➨ ➨

Interference reduction Possible coverage weakness in the short to medium distance range Traffic load distribution

ICN PLM CA NP

© SIEMENS Limited 1999

s

Omni vs. Sectorised OMNI cells - more difficult to optimise ➨

Electrical downtilt possible, however ✲

➨

same for entire cell

Parameters same for entire cell

Directional antennae ➨ ➨

narrower beam → easier to control interference tilting less efficient with wider beams Sectorised cell site with different downtilt angles

ICN PLM CA NP

© SIEMENS Limited 1999

s

Vertical Antenna Beam High gain antennae with sharp vertical lobe ➨

shadow under antenna

Ant. Effective height 0°

60 m

arctan

(60/4

00) = 8.

2° electrical d owntilt 3 dB-po int: 5.25 °

5°

City

400 m

Solution: Add mechanical downtilt ICN PLM CA NP

© SIEMENS Limited 1999

In practice: For cluttered environments reflections often compensate

s

Tilting Antenna downtilt often used to minimise interference ➨

Minimum: Vertical mail lobe pointing at cell edge

h

BS

➨

ICN PLM CA NP

Maximum: First null angle pointing at cell edge

© SIEMENS Limited 1999

s

Tilting Electrical vs. Mechanical downtilt 0° Mechanical

Electrical ➨

Advantages: ✲ ✲

➨

Better back lobe characteristics Better lower side lobe characteristics

Disadvantages: ✲

ICN PLM CA NP

0°

Antennas are more expensive

© SIEMENS Limited 1999

A combination of mechanical / electrical downtilt may be used

s

Tilting No Tilt

ICN PLM CA NP

Down Tilted 4 degrees

© SIEMENS Limited 1999

s

Antenna Diversity Type Rx ant.

Typical > 10

Dual polarisation Rx ant. 2

Rx ant. 1

Space diversity

Horisontal / vertical ➨ ➨ ➨ ➨

vertical polarisation in general good performance requires extra antenna for diversity

➨

➨

mobile antenna normally not held vertically when signals are reflected polarisation change (vertical normally dominates) cross polarised preferred ✲

➨

© SIEMENS Limited 1999

good performance in urban areas

save one antenna ✲

ICN PLM CA NP

Cross polarised

easier installation

s

Verification of RF Network Design Site check Site physical configuration evaluation Site-to-site distances and distribution Special features for improving coverage Site database configuration evaluation ➨ ➨ ➨

Tx power power control settings etc.

ICN PLM CA NP

BTS

© SIEMENS Limited 1999

s

Site Check Verify that site is implemented according to plan Check installation e.g. ➨ ➨ ➨ ➨ ➨

antenna spacing (diversity, isolation) antennae in one sector are installed in the same plane antennae alignment omni antenna installation cable installation Horisontal spacing

Vertical spacing

Rx

Tx

Antennas mounted in different planes

Omni

Tx k1 k2

k2

Rxd

Rx

Alignment of antennas

Rx

k

d

Tx

a= max 15 ° a

ICN PLM CA NP

Rx

a

d

© SIEMENS Limited 1999

Tx

Rxd

d

d

s

Antenna Isolation Isolation by vertical or horizontal separation between two antennas K73316.. A Horizontal

Horizontal

2250

2000

1750

1500

1250

1000

750

Vertical

500

Isolation /dB

60 50 40 30 20 10 0

A

Spacing A/ mm Source: Kathrein

ICN PLM CA NP

Vertical © SIEMENS Limited 1999

s

Antenna Isolation

A

60 40

Horizontal

20

Vertical

1250

1150

1000

900

750

650

500

0 400

Is olation /dB

Isolation by vertical or horizontal separation between two antennas K73416..

Horizontal A

Spacing A/mm Source: Kathrein

ICN PLM CA NP

Vertical © SIEMENS Limited 1999

s

Site Physical Configuration Antenna height ➨

➨

ideally sites within a given area classification should have similar heights if traffic distribution is uniform evaluate site height in terms of objective ✲ ✲ ✲

macrocell / minicell / microcell limitation of interference clear obstructions

Antenna tilt / directions ➨ ➨ ➨

avoid coverage gaps target priority areas limit interference

Appropriate antenna types ➨

sectorise omni cells?

ICN PLM CA NP

© SIEMENS Limited 1999

s

Site-to-Site Distances and Distribution For an area of uniform structure / terrain / traffic ➨

site-to-site distance should be uniform (assuming uniform site design)

Site distribution should reflect ➨ ➨

coverage characteristics / requirements capacity requirements

Typical case ➨ ➨ ➨

Downtown: High site density Suburban area: less dense Roads: Sites located along a line

ICN PLM CA NP

© SIEMENS Limited 1999

s

Special Features for Improving Coverage Microcell

➨ ➨ ➨

Repeaters ➨

➨ ➨ ➨ ➨

HCS, e.g.

alternative to microcell where the traffic needs are low indoor outdoor road coverage “coverage hole fill solution”

➨ ➨

large cells for car-coverage small cells for pedestrians Micro - Cell Site -Location Macro - Cell Site -Location ce Pla

Building Outlines

cro Ma

Building Outlines

Ce ll B ord

er

Scale = 0.5 Km

ICN PLM CA NP

distributed anteanne fibre optic repeater leaky cable

© SIEMENS Limited 1999

S tr e et

➨

for indoor coverage outdoor coverage in high capacity areas

Str eet

➨

Other indoor coverage solutions

ing ild s Bu tline u O

s

Cell splitting, Sectorisation Change from large cells to small cells Difficult , Expensive Mainly driven by capacity requirements Result: Improved indoor coverage

ICN PLM CA NP

© SIEMENS Limited 1999

s

DTM Check DTM resolution ➨

horisontal ✲

✲ ➨

macrocell (typical 50-100 m for roads, 50 m for small cities, 20 - 40 m for large cities) microcell (very high resolution, down to building level)

vertical - should be high

Source data ➨ ➨

heights and clutter derived from paper maps clutter and / or vector updates by satellite photographs / aerial photos for metropolitan areas

ICN PLM CA NP

© SIEMENS Limited 1999

s

Propagation Model Verification Wrong model wrong coverage prediction In general, standard models have high performance Highly specialised model may only be valid for a small area Model performance depends on accuracy of DTM To tune the model ➨ ➨ ➨

field strength measurements check existing model against measurements modify model parameters

ICN PLM CA NP

© SIEMENS Limited 1999

s

Link Budget Analysis Check for link budget imbalance downlink

Uplink Power Budget - Downlink Power Budget = 0! Link Power Budget is balanced! PA output pow er

downlink

com biner loss cable loss dow nlink Rx Sensitivity M S

B

ed nc a l a

Po

w

Bu er

et dg

Rx Sensitivity BS

cable loss uplink

antenna diversity gain

uplink M S Peak Pow er

ICN PLM CA NP

© SIEMENS Limited 1999

uplink

BTS

s

Link Budget Analysis

Uplink Power Budget - Downlink Power Budget Link Power Budget is unbalanced!

0!

RxLev/dBm -55,00

35% Coverage Loss @ 3dB! -65,00 RxLev for Indoor Coverage(90%) Links balanced -75,00 3dB unbalanced

55% Coverage Loss @ 6 dB!

6dB unbalanced

Caused by wrong assumption for Receiver Sensitivity Diversity Gain Propagation Environment Link Balancing via Optimization of Diversity Tower mounted amplifier High power amplifier

-85,00 0,20

ICN PLM CA NP

0,40 0,60 Distance from BTS in km

0,80

© SIEMENS Limited 1999

s

Link Budget Analysis Increasing BS Output? ➨

Unbalanced link budget

Better BS Rx sensitivity or pre-amplifier ➨

Must be matched by higher BS TX power for balanced link budget -110 dBm Uplink -107 Downlink

40 dBm 37 dBm ICN PLM CA NP

© SIEMENS Limited 1999

dBm

s

Dropped Call Analysis How to measure ➨

drive tests ✲ ✲

➨

repeated call setups (preferred) continuous calls

OMC measurements

Reasons for dropped calls ➨ ➨ ➨ ➨ ➨

lack of coverage interference problems handover problems lack of synchronisation in network problems with other parts of the network

ICN PLM CA NP

© SIEMENS Limited 1999

s

Call Setup Analysis How to measure ➨

drive tests ✲

➨

repeated call setups

OMC measurements

Reasons for failed call setups ➨ ➨

lack of coverage database problems ✲ ✲

database inconsistencies parameter settings, e.g. – RXLEV_ACCESS_MIN, RACHBT, RACH_MAX_RETRANS – cell reselection related parameters

➨

network congestion

ICN PLM CA NP

© SIEMENS Limited 1999

s

MAXRETR Slotted ALOHA mechanism: Several users may attempt to access channel simultaneously ➨ ➨

in case of collision new attempts are made MAXRETR: Maximum no. of retries allowed

H(1) RAC (2 ) H C RA H AGC

RA CH

BTS

MS ✲ ICN PLM CA NP

E.g: MAXRETR = 2 © SIEMENS Limited 1999

s

Handover Performance Analysis When moving from one cell to another (neighbour cells) handovers are necessary SIEMENS AG MON MAR15 15:18:41

Too many neighbours

SCALE 1:2500

Inaccurate handover decision

EqualPowerBoundary Mutual Neighbour Non-Mutual Neighbour Missing Neighbour Too many Neighbours

Handover Failure & Dropped Call

Missing Neighbour definition

Handover Failure

Dropped Call

ICN PLM CA NP

© SIEMENS Limited 1999

s

Handover Parameters Objectives: ➨

➨

mobile should be connected to the “best”cell avoid unnecessary handovers

Consequence ➨ ➨

ICN PLM CA NP

good speech quality less dropped calls

© SIEMENS Limited 1999

s

Consequence of Missing Neighbours Defined neighbours Server Missing neighbour Interferer

f1 f1

Missing neighbour cells

Cell dragging

Cell dragging

Poor RxQual

Poor RxLev

Interference

Dropped Calls

ICN PLM CA NP

Congestion

© SIEMENS Limited 1999

Exceeded distance

Poor PBGT

s

Consequence of Many Neighbour Definitions Only about 100 measurement samples are possible during one measurement period for all defined neighbour cells Number of BCCH carriers In BCCH Allocation 32 16 10 8 :

Number of samples per Carrier in SACCH multiframe 3-4 6-7 10-11 12-13 : (Rec. GSM 0508)

Too many neighbour cells Inaccurate signal level measurement False handover decisions Dropped Calls ICN PLM CA NP

© SIEMENS Limited 1999

Problem: Sites with too large coverage area

s

Handover Measurements Handover due to a better cell (RxLev_1 > RxLev_Full)

Handover due to bad quality Can also be analysed by statistics ICN PLM CA NP

© SIEMENS Limited 1999

s

Handover Parameters Fine-tuning of handover parameters ➨

Moving cell boundaries in order to ✲ ✲ ✲

➨

➨

Enhance success rate for critical handovers Minimise local interference at the cell edge Traffic load sharing between cells

Compared to other opimisation measures improvement potential is limited Affected by ✲ ✲

ICN PLM CA NP

Measurement averaging Power control parameters

© SIEMENS Limited 1999

PS! Neighbours should in general be mutual

s

Radio Link Measurements BTS measurements (Uplink): ➨ ➨ ➨ ➨

Signal level Quality BS-MS distance (Interference levels in idle time slots)

BSC

UL DL Neighbour

ICN PLM CA NP

MS

© SIEMENS Limited 1999

BTS

s

Radio Link Measurements MS measurements (Downlink) ➨ ➨ ➨

Signal Level Quality Signal levels of neighbouring cells (BCCH) ✲

BSC

Strongest 6 are reported to the Network

UL DL Neighbour ICN PLM CA NP

MS © SIEMENS Limited 1999

BTS

s

Radio Link Measurements BSC (In general) ➨

Collects all data ✲ ✲

BTS and MS send measurement reports every 480 ms Makes handover decisions

BSC

Siemens Network, BTS makes HO decisions

UL DL Neighbour ICN PLM CA NP

MS © SIEMENS Limited 1999

BTS

s

Radio Link Measurements Radio link measurements averaging ➨

BTS (BSC) receives measurement samples from BTS + MS ✲

➨

every SACCH-Multiframe (480ms,104 TDMA frames)

“Gliding Window” ✲ ✲

averaging Window size (max.31) Window is cleared after call setup or handover

32 27 23 29 29 21 19 22 23 21 Average value = 24

ICN PLM CA NP

© SIEMENS Limited 1999

s

Radio Link Measurements F

F

S

S

F

F

F

S

S

F

32 27 23 29 29 21 19 22 23 21

Measurement Values each SACCH Multiframe (0.48s)

32 32 27 27 23 29 29 29 21 21 Average value = 27 – W_Lev_Full = 2 – W_Lev_SUB = 1 – Gliding Window = 5

ICN PLM CA NP

© SIEMENS Limited 1999

s

Handover Algorithm Handover Handover Decision Decision

IRQUAL

yes

Inter-cell HO due to Quality

no LEV

yes

Inter-cell HO due to Level

yes IAQUAL no

yes

Inter-cell HO due to Distance

no

ICN PLM CA NP

PBGT

Inter-cell HO Power Budget

no

no DIST

yes

© SIEMENS Limited 1999

No handover action

Intra-cell HO due to Quality

s

Handover Criteria Handover Region (due to quality and level) Rx_Qual

L_Rx_Lev_XX_IH

7 Intracell HO due to Quality

Intercell HO due to quality

L_Rx_Qual_XX_H No handover action due to quality or level

Intercell HO due to level

0 ICN PLM CA NP

L_Rx_Lev_XX_H © SIEMENS Limited 1999

63

Rx_Lev

s

Handover Decision Handover Types Intercell HO due to Quality

Decision Criteria 1. RXQUAL_XX > L_RXQUAL_XX_H 2. RXLEV_XX < L_RXLEV_XX_IH 3. XX_TXPWR = Min (XX_TXPWR_Max,P) HO due to Level 1. RXLEV_XX > L_RXLEV_XX_H 2. XX_TXPWR = Min(XX_TXPWR_Max,P) HO due to Distance 1. MS_BS_DIST > MS_Range_Max HO due to 1. RXLEV_NCELL(n) > RXLEV_MIN(n) Power Budget + Max (0,MS_TXPWR_MAX(n)-P) 2. PBGT(n) > HO_MARGIN(n) Intracell HO 1. RXQUAL_XX > L_RXQUAL_XX_H due to Quality 2. RXLEV_XX > L_RXLEV_XX_IH

ICN PLM CA NP

© SIEMENS Limited 1999

s

Intracell Handover Stay within cell, change frequency / time slot situation ➨ ➨

in general interference different on different timeslots change to a different cell may be unnecessary

Interferer: f1

0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7

Sever: f1 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7

➨ ➨

higher traffic load higher likelihood on other timeslots not effective with frequency hopping ✲

ICN PLM CA NP

parameter settings for intracell handover should be set to reduce such handovers © SIEMENS Limited 1999

s

Intracell Handover Check for simultaneous occurrence of: ➨ ➨

Poor quality (high Rx_Qual) Sufficient signal level ✲

L_Rx_Lev_XX_IH Rx_Qual

L_Rx_Lev_XX_IH Intracell HO due to Quality

L_Rx_Qual_XX_H

L_Rx_Lev_XX_H

ICN PLM CA NP

© SIEMENS Limited 1999

Rx_Lev

s

Level Handovers Adjacent cell not stronger than current cell + HO margin Serving cell has insufficient coverage “emergency handover” to cell with better coverage Rx_Lev

➨

Server

HOMARGIN

HO_Threshold_Lev

neighbour MinHOReqInt

ICN PLM CA NP

© SIEMENS Limited 1999

Driven route

s

Level Handovers Receiver limit sensitivity L_RXLEV_XX_H (outgoing level HO) L_RXLEV_XX_IH (inter HO / intracell quality HO) RXLEV_MIN (incoming HO)

BTS

ICN PLM CA NP

RXLEV_MIN threshold for cell to accept incoming handover L_RXLEV_XX_H threshold for initiating outgoing handover due to signal level relation with RXLEV_MIN will determine hysteresis L_RXLEV_XX_IH threshold for initiating inter / intracell quality HO © SIEMENS Limited 1999

s

Distance Handover Maximum allowable BS-MS distance ➨

Default: MS_Range_Max=61 (bits Timing Advance,TA) ✲

Maximum value: 63, corresponding to 35 km

G S M : ma

➨

Enhanced by “Extended Cell”

Normally used in combination with other criteria, e.g. ✲ ✲

ICN PLM CA NP

x 35 km

cross-water propagation, elevated bridges etc.

© SIEMENS Limited 1999

s

Power Budget Handover Select cell with better signal level at given location HO margin ➨ ➨

Large enough to avoid “ping-pong HO” small enough to allow fast HO

BTS2

BTS1

Ping-Pong HO

1. RXLEV_NCELL(n) > RXLEV_MIN(n) + Max(0,MS_TXPWR_MAX(n)-P) 2. PBGT(n) = RXLEV_NCELL(n)-(RXLEV_DL+PWR_C_D) +Min(MS_TWPWR_MAX(n),P)-Min(MS_TXPWR_MAX(n),P) > HO_MARGIN(n) ICN PLM CA NP

© SIEMENS Limited 1999

s

Cell Reselection C1-criterion for cell access: C1 C1==AV_RXLEV AV_RXLEV- -RXLEV_ACCESS_MIN RXLEV_ACCESS_MIN- -MAX(0,MS_TXPWR_MAX_CCH-P) MAX(0,MS_TXPWR_MAX_CCH-P)>>00

✲

MS takes 5 samples of the received level on each RF carrier which

are averaged AV_RXLEV = 1/5 * (RXLEV1+RXLEV2+…+RXLEV5)

ICN PLM CA NP

© SIEMENS Limited 1999

s

Cell Reselection DL

For example:

BTS MS MS class 5 (GSM900) AV_RXLEV=-97 dBm ➨ ➨

RXLEV_ACCESS_MIN = -100 dBm MS_TXPWR_MAX_CCH = 29 dBm (0.8W) C1 = -97 - (-100) - Max(0,33-29) = -1

ICN PLM CA NP

© SIEMENS Limited 1999

s

Cell Reselection For example:

DL

MS class 5 (GSM900)

Operator B BTS

DL ✲

Operator A BTS ✲ ✲

MS

✲

RXLEV_ACCESS_MIN = -100 dBm MS_TXPWR_MAX_CCH = 33 dBm (2W)

RXLEV_ACCESS_MIN = -110 dBm MS_TXPWR_MAX_CCH = 33 dBm (2W)

•MS receives signal from Operator A and B = -90 dBm Operator A Operator B C1 = -90 - (-110) - Max(0,33-29) = +16

ICN PLM CA NP

C1 = -90 - (-100) - Max(0,33-29)

✓

= +6

© SIEMENS Limited 1999

✗

s

Cell Reselection C1 criteria ➨

Same Location Area ✲

➨

C1 (neighbour cell) > C1 (serving cell)

Different Location Area ✲

C1

C1 (neighbour cell) > C1 (serving cell) + Cell_Reselect_Hysteresis High power class MS Low power class MS Cell_Reselect_Hysteresis

BTS1

ICN PLM CA NP

BTS2

© SIEMENS Limited 1999

s

Speech Quality Analysis Parameters ➨ ➨ ➨

Causes of interference

RxQual Frame Erasure Rate (FER) Speech Quality Index (SQI)

➨ ➨ ➨

co-channel interference adjacent channel interference intermodulation ✲

Measurements ➨

✲ ➨

➨

Drive test

multipath interference

preferably continuous call

OMC statistics

Cause for poor quality ➨

➨ ➨

mainly on one link only

low signal strength (coverage related interference low signal strength and interference

ICN PLM CA NP

Interfering cell of base station within GSM network

Base station within GSM – Network

© SIEMENS Limited 1999

s

Downlink Interference Measurement Typical requirement ➨ ➨

speech: RxQual 4 data: RxQual 3 BER % 0.0 - 0.2 0.2 - 0.4 0.4 - 0.8 0.8 - 1.6 1.6 - 3.2 3.2 - 6.4 6.4 - 12.8 > 12.8

RxQual 0 1 2 3 4 5 6 7

With frequency hopping: RxQual not a valid parameter ICN PLM CA NP

© SIEMENS Limited 1999

s

Frequency Changes Sometimes necessary to minimise interference As network reaches capacity limit this becomes difficult ➨

Other frequencies may be affected by the change

Can be done at either interfering cell or victim cell ➨

Choice: Whichever happens to be easier to change

Existing plan may be entered into planning tool as “constraints” ➨

search for “optimum” frequency allocation for a given cell

At a certain point the whole network e.g. in a city may have to be re-planned

ICN PLM CA NP

© SIEMENS Limited 1999

s

Frequency Changes BCCH/TCH swapping ➨

Method sometimes used: Alternate between clusters

BCCH: 794 TCH:797

BCCH: 794

after

before

➨ ➨ ➨

Effectiveness depends on TCH traffic load BCCH / TCH sub-bands are mixed Could be used as a temporary measure ✲

ICN PLM CA NP

BCCH: 797 TCH:794

BCCH: 794

while traffic load is low

© SIEMENS Limited 1999

s

BSIC Optimisation Base Station Identity Codes ➨

Used by the MS to distinguish between cells using the same frequency ✲

Co-Channel cells must have different BSICs

f9 f9

f9

ICN PLM CA NP

© SIEMENS Limited 1999

s

Call Setup/Handover Mechanisms 20-25 dB street corner loss: Fast handovers required ✲ ✲

➨ ➨ ➨ ➨

Micro-micro Micro-macro

Fast measurement averaging Carefully tuned handover thresholds Small handover margins Short penalty timers

ICN PLM CA NP

© SIEMENS Limited 1999

s

Location Area Codes Purpose ➨ ➨

identify location area in incoming call is paged to all BTS’s within LA

Large location area ➨ ➨

advantage: less location updates (reduced SDCCH load) disadvantage: more paging traffic

Boundaries should not cross high traffic areas Cell reselection across LA boundaries ➨

Parameter Cell_Reselect_Hysteresis (typ. 4 dB) used to avoid unnecessary signalling due to ping-pong cell reselections

ICN PLM CA NP

© SIEMENS Limited 1999

s

Interference Reduction Power Control Frequency Hopping Discontinuous Transmission DTX

ICN PLM CA NP

© SIEMENS Limited 1999

s

Power Control Quality-triggered PC ➨

e.g. L_RXQUAL_XX_P = 4 ✲

➨

Triggers a power increase at poor quality

e.g. U_RXQUAL_XX_P = 1 ✲ ✲ ✲

Triggers a power reduction at good quality Virtually disabled by setting to “highest” RXQUAL value Level criterion is more suitable for power reduction

Level-triggered PC ➨

e.g. L_RXLEV_XX_P = 25 (-85 dBm) ✲

➨

e.g. U_RXQUAL_XX_P = 35 (-75 dBm) ✲

ICN PLM CA NP

Triggers a power increase at bad level Triggers a power reduction at good level © SIEMENS Limited 1999

s

Power Control RXQUAL

Power Increase (bad quality) L_RXQUAL_XX_P Power Decrease (Good Level)

Power Increase (bad level) L_RXQUAL_XX_P

Power Decrease (good quality) RXLEV L_RXLEV_XX_P

U_RXLEV_XX_P

2*POW_RED_STEP_SIZE ICN PLM CA NP

© SIEMENS Limited 1999

s

Frequency Hopping Cyclic / Pseudo Random hopping Baseband / Synthesized hopping

0

SDCCH

1

Call 1

5

Call 2

f1

0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7

f2

0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7

f3

0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 TDMA frame (8 time slots)

ICN PLM CA NP

0 BCCH

© SIEMENS Limited 1999

s

Frequency Hopping Cyclic / Pseudo Random hopping Baseband / Synthesized hopping

0 BCCH 0

SDCCH Call 1 Call 2

f1

0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7

f2

0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7

f3

0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 TDMA frame (8 time slots)

ICN PLM CA NP

© SIEMENS Limited 1999

s

Frequency Hopping Cyclic / Pseudo Random hopping Baseband / Synthesized hopping

0 BCCH 0

SDCCH

1

Call 1

5

Call 2

f1 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7

f2,f3,f4,f5,f6,f7

f6 f3 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7

f2,f3,f4,f5,f6,f7

f4

f7

0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 TDMA frame (8 time slots) ICN PLM CA NP

© SIEMENS Limited 1999

s

DTX Goal: Reduce speech data rate from 13 kbps (user speaking) to 500 bps (enough to encode background noise) ➨ ➨

reduce MS power consumption reduce the interference in a cell

SBS parameter for DTX / VAS administration ➨

DTXUL ->

➨

DTXDL ->

0 : MS may use DTX (If possible) 1 : MS shall use DTX 2 : MS shall not use DTX FALSE : downlink DTX disabled at BTS TRUE : downlink DTX enabled at BTS

PS! No gain for data communications ICN PLM CA NP

© SIEMENS Limited 1999

s

Channel Configuration Channel Type TCHFULL MAINBCCH MBCCHC SDCCH TCHF&HLF* BCBCH* SCBCH* CCCH* ✲

Channel Combination TCH/F + FACCH/F + SACCH/F FCCH + SCH + BCCH + CCCH (AGCH+PCH+RACH) FCCH + SCH + BCCH + CCCH + 4 (SDCCH+SACCH) 8 (SDCCH + SACCH) TCH/H(0) + FACCH/H (0) + SACCH/H(0) + TCH/H(1) FCCH + SCH + BCCH + CCCH + 3 (SDCCH+SACCH) + CBCH 7 (SDCCH + SACCH) + CBCH BCCH + CCCH

For example,

Note: * in SBS BR 3.0

– 1TRX : TS0 -> BCBCH TS1-7 -> TCHFULL – 2 TRXs : TRX0, TS0 -> MAINBCCH TRX0, TS1 -> SCBCH TRX0, TS2-7 -> TCHFULL TRX1, TS0-7 -> TCHFULL ICN PLM CA NP

© SIEMENS Limited 1999

s

Capacity Enhancements Easy approach: Add TRX’s Problem: No more frequencies: ➨

Options ✲ ✲ ✲ ✲ ✲

Traffic load distribution Interference optimisation features: FH, PC, DTX Sectorisation: Increasing cell density Cell splitting: Increasing site density HCS – – – –

✲ ✲ ICN PLM CA NP

Dual band operation (e.g. GSM900/DCS1800) Dual mode operation (e.g. GSM900/DECT) Underlay / Overlay Overlaid micro- and picocells

Half rate coding Migration to 3rd Generation Systems © SIEMENS Limited 1999

s

Adding TRX Congested cells found by OMC measurements Sec TRX GOS 2% Week1 Week2 Week3 Week4 Week5 Week6 Week7 BTS1 1 3 14.9 10.53 9.66 10.21 9.88 10.54 9.97 10.37 BTS2 2 2 8.2 7.43 7.26 7.59 6.98 7.55 8.02 8.33 BTS3 3 3 14.9 11.92 11.4 12.12 11.82 11.75 12.02 12.15 ➨ ➨

Sector 2 will experience congestion Sometimes percentage limit, e.g. 80%, of full load defined ✲

Sector 3 is near that limit

Possible limitations of TRX extensions: ➨

Need for changed hardware configuration costly ✲

➨

ICN PLM CA NP

e.g. new BTS rack needed

Frequency Spectrum limited © SIEMENS Limited 1999

s

Interference Reduction Features Frequency Hopping (FH) Dynamic Power Control (PC) Discontinuous Transmission (DTX) ➨

allow tighter frequency re-us (already considered for 40-60 Erl./km2 in macrocell layer with 5 to 10 MHz)

✓ ✓No Noadditional additionalsites sitesor orfrequencies frequenciesrequired required ✓ ✓Available, Available,stable stable ✓ ✓Implementation Implementationcauses causesno nodisturbance disturbanceofof network networkoperation operation ✗✗LLittle ittleor orno noeffect effectififavailable availablespectrum spectrumisisvery very limited limited(BCCH (BCCHlimitations) limitations) ICN PLM CA NP

© SIEMENS Limited 1999

s

Traffic Load Distribution Traffic in a cell related to cell coverage area If sufficient overlap between cells: ➨

reduce traffic by changing cell boundary ✲ ✲ ✲

➨

antenna downtilt reduce power (PWRRED) alter handover boundaries

Usually a temporary solution only

Default HO boundaries Changed HO boundaries

ICN PLM CA NP

© SIEMENS Limited 1999

s

Call Setup/handover mechanisms Relieve macrocells from traffic

➨ ➨

Umbrella type handover into microcells “Directed retry” ✲

ICN PLM CA NP

Allows call setup In second-best server, shares traffic resources between layers

© SIEMENS Limited 1999

s

Hierarchical Cell Structures Underlay/Overlay Umbrella cells: Dominant site with large coverage area ➨

low traffic - fast mobiles

Macrocells: Antenna above average rooftop level ➨

normal traffic

Microcell: Antenna below average rooftop level ➨

cover small high traffic areas

Picocell: Antenna usually indoors ➨

coverage to building or parts thereof - e.g. Business users

ICN PLM CA NP

P

C

I

E

C

L

O

Ls

Indoor coverage Outdoor Installatio

Hotspot n

Parking lot

Contiguous Microcellular Coverage

Subway Coverage Extension

© SIEMENS Limited 1999

s

Concentric cells C/I = 17 dB Signal

f3

level

C/I = 17 dB level Signal

C/I = 0 dB

f1

f2

f3

“Inner cell” can use 1 x 3 reuse pattern Special handover mechanisms between layers Limited gains for uniform traffic distribution

ICN PLM CA NP

© SIEMENS Limited 1999

s

Overlaid Micro- and Picocells The smallest cells should absorb most of the traffic in their coverage area Larger cells for fast moving mobiles / areas not covered by small cells

Macrocells Microcells Picocells ICN PLM CA NP

© SIEMENS Limited 1999

s

Microcell Frequency Planning Different resolutions required for different layers ➨

flexibility of planning tool needed

Dedicated frequency bands for different layers ➨ ➨

Reduce complexity of frequency optimisation task Guard band may be needed to avoid adjacent channel interference

Call Setup/handover strategy ➨ ➨ ➨

Serving BTS

reduce macrocell traffic determine mobile speed Fast handovers ✲

Loss around street corner: 20 dB! Micro BTS

ICN PLM CA NP

© SIEMENS Limited 1999

s

Speed Sensitive Handovers Mechanisms to separate fast from slow mobiles ➨

mobile class ✲

➨

measurement of the timing advance delta ✲

➨

only works for direction away from site

cell type ✲

➨

today mostly same class is used (e.g. GSM900 class 4)

try to keep handovers within same layer unless speed change

mean time between handovers

ICN PLM CA NP

© SIEMENS Limited 1999

s

Half Rate Coding / Dual Rate Operation Has potential to double network capacity ➨

Advantages: ✲ ✲

➨

No additional sites / frequencies required Minimum investment for infrastructure upgrade

Disadvantage: ✲

Speech quality degradation (reduction of speech bit rate from 13 kb/s to 6.5 kb/s) – Especially mobile-to-mobile calls

➨

Gain depends on ratio full rate users / half rate users / data traffic

ICN PLM CA NP

© SIEMENS Limited 1999

s

Cell Parameter Optimisation Default parameter sets: ➨ ➨

PS! Standard setting suitable for most cases Starting point for possible optimisation, however ✲

➨

more relevant after other optimisation activities

Different parameter standards may be used for ✲ ✲ ✲

different area types BTS types etc.

Danger ➨

many parameters ✲ ✲

ICN PLM CA NP

easy to lose overview

inconsistencies deterioration of quality

© SIEMENS Limited 1999

s

Effect

Possible Network Optimisation Measures

Dual mode

Fine tuning of antenna orientation and tilt Adding TRX

Repeaters Preamps

Frequency Changes

Sectorisation

HR Cell parameter setting

Underlay/ Overlay

FH, PC, DTX

Dual band

Cost, Effort

ICN PLM CA NP

© SIEMENS Limited 1999

Cell spiltting

Overlaid microcells

s

Increasing Network Capacity The relationship between quality and capacity ➨

In a congested network, quality can deteriorate very quickly: Congestion Poor speech quality

Extended call setup times Interference/ Noise

Dropped call

Unavailability of service

– Violation of all 4 basic quality criteria ICN PLM CA NP

© SIEMENS Limited 1999