30563107 Planning and Optimization of FH in Gsm

s

Network Design and Consulting

Planning & Optimisation of

Frequency Hopping in

GSM Networks

Authors:

U. Rehfuess, ICM N MRDr. K. Dietrich, ICM N MRA. Volke, ICM N MRB. Kronmueller, ICM N ST

s


Outline

Planning & Optimisation of Frequency Hopping in GSM Networks

Implementation Aspects

Frequency Assignment in FH Networks

BSS - Database Parameters

Optimisation Aspects

Summary

s


Implementation Aspects


s


Frame N° 0Frame N° 1Frame N° 2Frame N° 3

Baseband FH

RF1

RF2

RF3

RF4

BB1

BB2

BB3

BB4

Logical Channel

1

2

3

4

Synthesizer FH

• Mobiles use Synthesizer Hopping only• BS implementation: power down, synthesizer re-

tuning and power up again within guard period 2 Synthesizers are implemented

RF1..n

RF1..n

RF1..n

BB1

BB2

BB3

BB4

RF1..n

Logical Channel

1

2

3

4

Implementation AspectsKey Differences Between Baseband and Synthesizer FH

s


Baseband FH Synthesizer FH

Implementation Aspects Combining Equipment in Baseband and Synthesizer FH

RF1

RF2

RF3

RF4

BB1

BB2

BB3

BB4

1

2

3

4

Filt

er

Com

bin

ing

TX Antenna

RF1..n

RF1..n

RF1..n

BB1

BB2

BB3

BB4

RF1..n

1

2

3

4

Hyb

rid C

om

bin

g

TX Antenna

• Narrow Band• Low insertion loss (3-4 dB)

• Wide band• Higher insertion losses (~3 dB/stage)• On-air combining possible (DUCOM)

s


Implementation Aspects Combiner, Link Budget of Synthesized Frequency Hopping

Power Output Power

Amplifier W

PA 25 W40 W

HPA 60 W

Combiner Attenuation

Type dB

DUCOM 2:1 2.5

DUCOM 4:1 5.7

HYCOM 1:1 2.0

HYCOM 2:1 3.7

HYCOM 4:1 6.5

FICOM 2:1 2.4

FICOM 4:1 3.0

FICOM 6:1 3.3

DUCIT 2.8

Combiner Losses and Output Power

Example: GSM 900

• Determine configuration and hardware status before SFHimplementation

• Determine necessary upgrades(TPU, PA, Combiner)

• Actualize and check Link budgets

• Introduce further HPA wherenecessary

• Max. number of TRXs per celldepends on hardware configuration

s


BFH SFH

TPU X

TPU 2 X X

PA X X X*

HPA X X

* not all types of first generation power amplifiers are suitable for for SFH

Implementation Aspects TPU, PA for Synthesized Frequency Hopping

s


• Software Release: BR 3.7 or higher

• Cell Synchronization: up to 2/2/2 BS 6x/2xup to 8/8/8 BS 24x

• No. of Hopping Frequencies:max. 16 per cell (BFH incl. BCCH)max. 15 per cell (SFH w/o BCCH)max. 64 per cell with BR 6.0

• BS11: SFH only (BR 4.0)

Implementation Aspects Hardware and Software for Synthesized Frequency Hopping

s


• Baseband hopping Narrowband RFcombining sufficient

BCCH TRX except for TS0 may hop

• Synthesizer hopping Wideband RF combining required

One TRX per hopping frequency required!

• No. of RF = No. of TRX

• No. of RF > No. of TRX BCCH TRX must not hop

More hopping frequencies than TRXs feasible

Implementation Aspects Key Differences Between Baseband and Synthesizer FH

s


Implementation AspectsHardware Requirements: Repeater

Wideband Repeaters:

• Usable for SFH and BFH

• Careful implementation (amplification of signals

in the whole frequency band)

Channel selective Repeaters:

• Usable for BFH

• Number of frequencies is limited

• Usually not usable in tight reuse scenarios

s


Frequency Assignment in Hopping Networks


s


Frequency PlanningProcess Frequency Assignment

Split of BandBCCH - TCH

DedicatedCommon

Multiple Reuse

Planning of BoundariesHopping – Non Hopping

Available Spectrumfor Hopping

Frequencyassignment with

fixed reuse schemes

Reuse 1x3 Reuse 1x1 other

Tools Interference Table Separation Settings

Tool optimizedfrequencyassignment

MAIO and HSNPlanning

Cyclic HoppingRandom Hopping

DatabaseGeneration

Planning of AnchorFrequencies in SFH

• Guideline for RF-planners

• Focus on SFH planning and hopping TCH - carriers

• BCCH - carrier assignment: planning with tool is always recommended

• Planning must be adjusted to each individual network

s


Frequency Planning Common Band - Dedicated Band - Multiple Re-Use Patterns

5 hopping frequenciesPC on, DTX on

[%]

90%@FER2%

Dedicated Band 71.8%

Dedicated Band

15 BCCH carriers 28 TCH carriersCommon Band

59.7%

Common Bandtotal operator bandwidth 8.6 MHz = 43 carriers

43 carriers for both BCCH and TCH

MRP 54.3%

15 BCCH carriers 12 TCH + 9 TCH + 7 TCH carriers

Multiple Re-use Patterns (MRP)

Achievable System Load

s


Fixed reuse scheme to all hopping cells

possible reuses:3/9, 2/6, 1/3, 1/1

Tool supported frequency assignment based on interference matrix considering FH gains

Frequency PlanningFrequency groups - Tool supported planning

TCH 1 TCH 3

TCH 2

TCH 1 TCH 3

TCH 2

TCH 1 TCH 3

TCH 2

TCH 1 TCH 3

TCH 2

s


Frequency Planning Planning of BCCH

BCCH 3 BCCH 2

BCCH 1

BCCH 9 BCCH 8

BCCH 7

BCCH 6 BCCH 5

BCCH 4

BCCH 13 BCCH 11

BCCH 10

BCCHe.g. 4 x 12 Reuse

• Reliability

• Neighbor Measurements

• BSIC Decoding

• BCCH Frequency active at all timeslots in the downlink-> no interference averaging

s


Cluster 1/3

Channel

1, 4, 7, 10, ...

2, 5, 8, 11, ...

3, 6, 9, 12, ...

With a deliberately MAIO - assignment and identical HSN assignment to sectors you can avoid adjacent-channel interference between the sectors within one site

Frequency Planning Examples for frequency groups (I)

Co-channel interference is avoided by the frequency groups MAIO TRX1 TRX2 TRX3 ...

Sector 1 0 2 4 ...

Sector 2 1 3 5 ...

Sector 3 0 2 4 ...

Min # RF 6 12 18

TCH C TCH B

TCH A

TCH C TCH B

TCH A

TCH C TCH B

TCH A

TCH C TCH B

TCH A

TCH C TCH B

TCH A

TCH C TCH B

TCH ATCH A

TCH B

TCH C

s


Frequency Planning Examples for frequency groups (II)

• Each sector within a site uses a different Frequency Group

• No co-channel collisions between sectors of a site

• Synchronisation between the sectors and MAIO management avoid adjacent channel collisions

• Homogeneous network:no co-channel collisions between serving cell and all nearest neighbour cells

TCH uses each frequency onlypart of the time (e.g. 50%)

50% fractional load

TCH 1 TCH 3

TCH 2

TCH 1 TCH 3

TCH 2

TCH 3

TCH 2

TCH 1 TCH 3

TCH 2

TCH 1

s


Cluster 1/1

• All sectors same frequency group

• Identical HSN to sectors of one site

• MAIO assignment to avoid co- and adjacent channel interference

Frequency PlanningExamples for frequency groups (III)

MAIO TRX1 TRX2 TRX3 TRX4 ...

Sector 1 0 6 12 18 ...

Sector 2 2 8 14 20 ...

Sector 3 4 10 16 22 ...

Min # RF 6 12 18 24

s


Frequency PlanningExamples for frequency groups (IV)

• Each sector within a site uses the same frequency group

• Synchronisation between the sectors and MAIO management required to avoid co-channel collisions

• Homogeneous network:Co-channel collisions between serving cell and nearest neighbor

TCH TCH

TCH

TCH TCH

TCH

TCH TCH

TCH

TCH TCH

TCH

s


Frequency PlanningHopping Sequence Generation (I)

MAI = (FN + MAIO) modulo N if HSN = 0 (cyclic hopping)

GSM 05.02.

MAI ... Mobile Allocation Index (integer 1...N-1)

FN ... TDMA Frame Number (0... 26*51*2048-1 = 2 715 647)

MAIO ... Mobile Allocation Index Offset (0 ... N -1)

N ... Number of allocated frequencies

For example: (MAIO=0)

MA = 1,4,7,10,13,16,19,21,24,27,30,33,36,39,41

set of ARFCN numbers to be

used in the hopping sequence

N=15

1. burst FN = 0: MAI = (0 + 0) mod 15 = 0 ARFCN = 12. burst FN = 1: MAI = (1 + 0) mod 15 = 1 ARFCN = 4

14. burst FN = 14: MAI = (14 + 0) mod 15 = 14 ARFCN = 4115. burst FN = 15: MAI = (15 + 0) mod 15 = 0 ARFCN = 1 16. burst FN = 16: MAI = (16 + 0) mod 15 = 1 ARFCN = 4etc...

s


Frequency PlanningHopping Sequence Generation (II)

MAI = (S + MAIO) modulo N if HSN 0 (random hopping)

with:

S = M’ if M’ < N

S = (M’ + T’) modulo N else

M’ = M modulo [2^Integer(log2(N)+1)]

T’ = T3 modulo [2^Integer(log2(N)+1)]

M = T2 + RNTABLE((HSN xor T1R)+T3)

T1R, T2, T3 ... Different Time ParameterRNTABLE ... Table of 114 Integer numbers

s


Frequency PlanningExample for MAIO - Management (I)

Frequency group 1x1 reuse / Random Hopping (1, 2, 10, 7, . . . )

10

7

Time (TDMA - frame)

Time (TDMA - frame)

1

35

911

13

1517

2

46

8

TRX0

TRX1

TRX2

TRX3

BCCH

MAIO = 2

MAIO = 8

MAIO = 14

TRX0

TRX1

TRX2

BCCH

MAIO = 4

MAIO = 10

MAIO = 16 TRX3

TRX0

TRX1

TRX2

TRX3

BCCH

MAIO = 0

MAIO = 6

MAIO = 12

1012

14

1618

1 2 3 4 5 6 1817161514131211987 10

1214

16

182

4

68

7

9

. . .

. . .

. . .

. . .

. . .

. . .

. . .

. . .

. . .

Time (TDMA - frame)

s


Frequency PlanningExample for MAIO - Management (II)

Frequency group 1x1 reuse / Random Hopping (1, 2, 10, 7, . . . )

1 2 3 4 5 6 1817161514131211987 10

Avoid Co - channel collision:

min # RF = number of hopping TRX (example 9 frequencies)

Avoid Adjacent - channel collision:

only odd or even RF numbers on air at same time

Minimum total number of frequencies for hopping system with MAIO - Management = 2* number of hopping TRX of site (18 frequencies in example)

TRX0

TRX1

TRX2

TRX3

BCCH

MAIO = 2

MAIO = 8

MAIO = 14

TRX0

TRX1

TRX2

BCCH

MAIO = 4

MAIO = 10

MAIO = 16 TRX3

TRX0

TRX1

TRX2

TRX3

BCCH

MAIO = 0

MAIO = 6

MAIO = 12

s


Frequency PlanningExamples for frequency groups and MAIO - Assignment

TRX0

TRX1

TRX2

TRX0

TRX1

TRX2

TRX0

TRX1

TRX2

TRX0

TRX1

TRX2

TRX0

TRX1

TRX2

TRX0

TRX1

TRX2

TRX0

TRX1

TRX2

HSN = 1

HSN = 2

HSN = 3

TRX0

TRX1

TRX2

TRX3

f B

f B

f B

f B

f B

f B

f B

f C

f C

f C

f C

f C

f C

f C TRX3

TRX0

TRX1

TRX2

TRX3

BCCH

BCCHBCCH

BCCH

BCCHBCCH

BCCH BCCH

BCCH

MAIO = 0

MAIO = 1MAIO = 0

MAIO = 0

MAIO = 1MAIO = 0

MAIO = 0 MAIO = 1

MAIO = 0

MAIO = 2

MAIO = 3MAIO = 2

MAIO = 2

MAIO = 3MAIO = 2

MAIO = 2 MAIO = 3

MAIO = 2

f A

f A

f A

f A

f A

f A

f A

MAIO = 5MAIO = 4

MAIO = 4

Frequency group: A: 1 4 7 10 13 16B: 2 5 8 11 14 17C: 3 6 9 12 15 18

s


Frequency Planning ToolsThe Automatic Frequency Planning Process

Input datafrom radio networkplanning tool

•Automatized Planning Routines•Variety of Planning Algorithms

•Setting of planning constraints•Common / Dedicated Band Planning

•Global / Local Parameter Settings

Frequency Assignment

Live Network Data

Minimisation of

interference

•Consideration of FH, PC, DTX

Evaluation of the assignments•C/I and FER plots•C/I and FER analysis on per carrier basis

Evaluation of the assignments•C/I and FER plots•C/I and FER analysis on per carrier basis

s


Frequency Planning ToolsThe SIEMENS Advanced Automatic Frequency Planning Tool

Efficient algorithms for different optimization targets: Minimizing global interference

Minimizing worst interfering cell relations

...

Features for advanced network planning strategies Frequency hopping

Power Control

Discontinuous transmission

Graphical evaluation of frequency assignments based on C/I

FER

Very good results in European research program COST 259 benchmarks in quality of result at short execution times (typically seconds to minutes)

High performance proved in live networks with different customers

s


Frequency Planning Tools Consideration of Radio Link Control Options

Automatic consideration of hopping gains and

interference reduction due to PC and DTX on cell basis

during

• interference matrix calculation

• optimum assignment of frequencies by using highly

efficient optimisation algorithms

Graphical evaluation of the assignment results based

on FER

s


EIRP - Path

loss

EIRP - Path

loss

Potentially serving signal SC

Potentially interfering signal SI

Ai

Ai

Ai 50%

Frequency Planning ToolsGeneration of the interference matrix

s


Frequency Planning Tools

Required C/I in FH-GSM (TU3), Cyclic Hopping

FH Gains as determined via Real Network Simulations

Shift: 6.5 dB 13.5 dBGain: up to 7 dB

NH2 Ch3 Ch4 Ch5 Ch8 Ch

50%

s


Frequency AssignmentFrequency Reuse & C/I values (Non Hopping)

Required no. of frequencies Cluster size / Reuse distance: q = SQRT(3*N) C/I rule of thump: C/I abs 1,5 * N2

N Anzahl f q C/I [dB]2 6 2,45 7,783 9 3,00 11,304 12 3,46 13,805 15 3,87 15,746 18 4,24 17,327 21 4,58 18,668 24 4,90 19,829 27 5,20 20,8510 30 5,48 21,7612 36 6,00 23,3415 45 6,71 25,2818 54 7,35 26,8720 60 7,75 27,78

s


Frequency Planning Tools

Required C/I in FH-GSM for different environments

Typical frequency hopping gainThe following table shows the typical gain from frequency hopping in a GSM 900 network (example of the signal-to-noise ratio required to obtain 0.2% residual BER for class 1b bits):

Frequency hopping TU3 TU50 HT100

None 11.5 7.5 6.82 frequency 10.0 6.5 6.74 frequency 8.25 6.0 6.68 frequency 7.5 6.0 6.616 frequency 6.75 6.0 6.6

Source: SIEMENS TED-BSS

s


Frequency Planning Tools Analyses of FER

Graphical FER analysis of an SFH network

1x3 reuse, 0,3 fractional load 1x3 reuse, 0,6 fractional load

< 1%

< 2 %

< 3 %

3 %

FER in %

< 1%

< 2 %

< 3 %

3 %

FER in %

s


Frequency AssignmentExample for Tool-supported Planned Reuse (I)

3

33

4

443

22

3

43

3

33

4

22

4

44

4

443

33

2

44

2

44Network Example:

• 11 Sites

• 33 Cells• 6 cells 2 TRX• 12 cells 3 TRX• 15 cells 4 TRX• 33 TRX BCCH• 75 TRX TCH

No. of TRX

s


Frequency AssignmentExample for Tool-supported Planned Reuse (II)

Given Spectrum: 42 channels

12 frequencies for BCCH - TRX

30 frequencies for TCH - TRX (hopping)

Reuse of: 4 30/4 = 7.5 frequencies per cell in average

5 30/5 = 6 frequencies per cell in average

6 30/6 = 5 frequencies per cell in average

7 30/7 = 4.2 frequencies per cell in average

Network Example:• 11 Sites• 33 Cells• 6 cells 2 TRX• 12 cells 3 TRX• 15 cells 4 TRX

s


Frequency AssignmentExample for Tool-supported Planned Reuse (III)

Planning Rule: (example)

1 Hopping TRX 3 frequencies2 Hopping TRX 4 frequencies3 Hopping TRX 6 frequencies

No. of assigned

frequencies for FH

Frequency Reuse Factor:

156 / 33 = 4.7 frequ. / cell in average

30 frequ. / 4.7 frequ. per cell = 6.3

4 / 6

4 / 66 / 4

3 / 4

3 / 44 / 3

2 / 3

4 / 66 / 4

No. of TRX

Network Example:• 11 Sites• 33 Cells• 6 cells 2 TRX• 12 cells 3 TRX• 15 cells 4 TRX

s


Frequency AssignmentExample for Tool-supported Planned Reuse (IV)

Separations for hopping TCH:• Intra cell separation: 3

• Intra site separation: 1

• Neighbour separation: 1

Interference Matrix for hopping TCH:• co-channel: C/I curve 7 dB (50% probability)• adjacent channel: C/I curve -6 dB (50% probability)

MAIO and HSN:• HSN = 0 for all cells (cyclic hopping)• MAIO = 0 for TRX1 (TRX0 = BCCH)

• MAIO = 1 for TRX2• MAIO = 2 for TRX3 etc.

s


Frequency AssignmentExample for Tool-supported Planned Reuse (IV)

etc....

Example for a site list:Site Id Sector TRX BCCH f1 f2 f3 f4 f5 f6 MAIO HSN0001 1 0 2 - -

1 1 19 36 41 0 02 0 4 - -2 1 13 18 21 25 30 42 0 02 2 13 18 21 25 30 42 1 02 3 13 18 21 25 30 42 2 03 0 9 - -3 1 15 23 27 32 0 03 2 15 23 27 32 1 0

0002 1 0 12 - -1 1 22 26 34 40 0 01 2 22 26 34 40 1 02 0 8 - -2 1 16 19 28 0 0

s


Frequency PlanningStrategies Cyclic Hopping - Random Hopping

Cyclic hopping sequence {... f4, f0, f1, f2, f3, f4, f0, f1, f2, f3 ...}, MAIO 0Cyclic hopping sequence {... f1, f2, f3, f4, f0, f1, f2, f3, f4, f5 ...}, MAIO 2

F

r e

q

u

e n

c

y

TDMA frame

f0

f1

f2

f3

f4

Principle of Cyclic Hopping

• Optimum frequency Diversity• Sufficient Interference diversity by avoiding frequency groups• No Interference diversity using frequency groups

Random hopping sequence {... f1, f4, f2, f0, f0, f3, f0, f1, f2, f4, ...}, MAIO 0Random hopping sequence {... f3, f1, f4, f2, f2, f1, f2, f3, f4, f1, ...}, MAIO 2

F

r e

q

u

e n

c

y

TDMA frame

f0

f1

f2

f3

f4

Principle of Random Hopping

• Optimum interference diversity

• Less frequency diversity

s


Frequency PlanningPlanning of Anchor Frequencies

Each TRX must be assigned with a fixed frequency which belongs to the hopping frequency group of this TRX

In case of disabling FH the TRX transmit the anchor frequency

Tool Supported Planning of anchor frequencies

2 - 5 - 8 - 11 - 14 - 17

Example Frequency group 1x3 reuse: A: 1 4 7 10 13 16B: 2 5 8 11 14 17C: 3 6 9 12 15 18

TRX0

TRX1

BCCH

TRXFREQ = 183 - 6 - 9 - 12 - 15 - 18

TRX0

TRX1

TRX2

TRX3

BCCH

TRXFREQ = 7

TRXFREQ = 13

TRXFREQ = 4

MOBALLOC = 1- 4 -7- 10 - 13 - 16 MOBALLOC = 1 - 4 - 7 - 10 - 13 - 16 MOBALLOC = 1 - 4 - 7 - 10 - 13 - 16

TRX0

TRX1

BCCH

TRXFREQ = 2......

......

s


BSS - Database Parameters for Radio Control Features


s


SpecificationName

Object/Package

DB Name Range Meaning

CA BTS/PKGBTSB

CALL 0...1023&...&0...1023

Cell Allocation:list of all frequencies used in this cell except theBCCH-frequency

HOPP_MODE BTS/PKGBTSO

HOPMODE BBHOPSYNHOP

Flag indicates whether baseband or synthesizerhopping is used

FH_SYS_NUMBER

FHSY FHSY 1...10 Number to identify a frequency hopping system

HSN FHSY HSN 0...63 Hopping sequence number0: cyclic hopping1...63: random hopping

MA FHSY MOBALLOC 0...1023& ...&0...1023

Mobile allocation:list of frequencies within the FH system(maximum number of frequencies = 16).

FH_SYSTEM_ID CHAN FHSYID 0...10 Specifies the frequency hopping system (givenby FHSYN) to which a channel shall belong. (0:no hopping)

MAIO CHAN MAIO 0...63 Mobile allocation index offset: defining thestarting frequency (number in the MA frequencylist) for a hopping sequence at a certain framenumber FN, i.e. different channels using thesame FH system shall have different MAIO.

BTS_ISHOPPING BTS/PKGBTSO

HOPP TRUEFALSE

Flag to enable/disable FH within the BTS

Source: M. Doss, ICN CA CV D 22

Database ParametersAdministration of Database Parameters for FH (I)

s


Database ParametersAdministration of Database Parameters for FH(I)

Maximum No. Of Hopping FrequenciesTheoretical Limit: 16 Frequencies (BFH)Practical Limit: 15 Frequencies (SFH)

In case of SFH all Time Slots on BCCH TRX (BCCH TS, SDCCH TS as well as TCH TS) must not hop

In case of BFH the BCCH TS must not hop (FHSYID = 0)

s


L_RXQUAL_XX_P

RXLEV63

U_RXQUAL_XX_P

0

7

RXQUAL

POW_RED_STEP_SIZE

L_RXLEV_XX_P U_RXLEV_XX_P

Power Decrease(good quality)

Power Decrease(good level)

Power Increase(bad level)

Power Increase(bad quality)

Database ParametersAdministration of Database Parameters for Power Control (I)

s


Specification Name DB Name/Object Range Meaning

MS_TXPWR_MAX MSTXPMAX /BTSB

2...150...15

maximum TXPWR an MS may use in the serving cell

BS_TXPWR_RED PWRRED/ TRX

0...6 Static reduction of the TRX output power:BS_TXPWR_MAX = PBTS - 2 * PWRRED

POW_INCR_STEP_SIZE

PWRINCSS/ PWRC

DB2,DB4,DB6

Step size for power increase in dB

POW_RED_STEP_SIZE

PWREDSS/ PWRC

DB2,DB4 Step size for power reduction in dB

P_CONFIRM PWRCONF/PWRC

1...31 Maximum interval for waiting for a confirmation of the new transmitpower level. unit: 2 TSACCH

L_RXLEV_DL_PL_RXLEV_UL_P

LOWTLEVDLOWTLEVU

0...63 RXLEV threshold on downlink/uplink for power increase

U_RXLEV_DL_PU_RXLEV_UL_P

UPTLEVDUPTLEVU

0...63 RXLEV threshold on downlink/uplink for power decrease

L_RXQUAL_DL_PL_RXQUAL_UL_P

LOWTQUADLOWTQUAU

0...7 RXQUAL threshold on downlink/uplink for power increase

U_RXQUAL_DL_PU_RXQUAL_UL_P

UPTQUADUPTQUAU

0...7 RXQUAL threshold on downlink/uplink for power decrease

P_CON_INTERVAL PCONINT / PWRC 0...31 Minimum interval between changes of the RF transmit power level

Database ParametersAdministration of Database Parameters for Power Control (II)

s


Specification NameObject/Package

DB Name Meaning

DTX indicator uplink BTS/BTSO DTXUL 0: MS may use DTX (if possible)1: MS shall use DTX2: MS shall not use DTX

DTX indicatordownlink

BTS/BTSO DTXDL FALSE: downlink DTX disabled at BTSTRUE: downlink DTX enabled at BTS

Database ParametersAdministration of Discontinuous Transmission (DTX)

s


No Frequency Hopping

0

1

2

3

4

5

6

7

0,1 1 10 100

FER@90% [%]

RX

QU

AL

@9

0%

2% FER

Cyclic FH 2 Frequencies

0

1

2

3

4

5

6

7

0,1 1 10 100

FER@90% [%]

RX

QU

AL

@9

0%

2% FER


0

1

2

3

4

5

6

7

0,1 1 10 100

FER@90% [%]

RX

QU

AL

@9

0%

2% FER


0

1

2

3

4

5

6

7

0,1 1 10 100

FER@90% [%]

RX

QU

AL

@9

0%

2% FER

Database ParametersAdministration of Database Parameters for FH - RxQual

s


Database ParametersParameter Settings for Control Loops using RxQual Measurements

Handover: Default Default Default (No Hopping) SFH (> 8 HF) BFH

HOLTHQUDL: 4 (5*) 5 ... 6 4 ... 5HOLTHQUUL: 4 (5*) 5 ... 6 4 ... 5HOAVQUAL: 8-2 8-2 8-2

Power Control:LOWTQUAD: 3 (4*) 4 ... 5 3 ... 4LOWTQUAU: 3 (4*) 4 ... 5 3 ... 4UPTQUAD: 1 1 1 UPTQUAU: 1 1 1PAVRQUAL 4-1 4-1 4-1

*) as recommended by database planning

s


Database ParametersFlexible Configuration of different Hopping Modes

Flexible configuration of the Hopping Mode for each cell of a BSC via BSS parameter:

• No Hopping

• Baseband Hopping

• Synthesizer Hopping

Configuration of the first SDCCH on the BCCH TRX mandatory

s


Optimisation Aspects


s


Points of Examination

Impact of RLC on Network Quality

Impact of RLC on Network Capacity

Comparison BFH - SFH

Comparison SFH Tight Reuse Implementation Cases

Optimizing Radio Parameters

(Thresholds for Handover and Power Control)

Optimisation AspectsField Trials

s


Optimisation AspectsNetwork Measurements Performance Measurements:

• Performance and Quality Indicators- Dropped Call Rate- TCH Drop Rate (Loss of Connection, Loss during Handover)

• Handover Statistics (Inter, Intra, Causes, Failures)

• Uplink Interference Measurements on Idle TCH

• SDCCH Performance

Test Mobile Measurements:

• RxLev, RxQual, FER, SQI (Speech Quality)

Tracer Measurements: Abis Protocol Analyses

s


0.0%

1.0%

2.0%

3.0%

4.0%

5.0%

No Hopping 4.55% 1.23% 1.90%

SFH 1x1 4.39% 1.08% 1.59%

SFH 1x3 4.40% 1.04% 1.52%

Dropped Call Rate TCH Drop Rate Call Drop Rate BS

Optimisation Aspects Performance Measurements: Quality Indicators

Comparison of Drop Rates: Non Hopping - SFH 1x3 - SHF 1x1

~ 4 % Improvement

~ 14 % Improvement

~ 18 % Improvement

s


0.0%

20.0%

40.0%

60.0%

80.0%

100.0%

No Hopping S1 / SFH 1x3 SFH 1x3

Uplink Quality (Per) Downlink Quality (Per) Uplink Strengh (Per) Downlink Strength (Per)

Distance (Per) Better Cell (Per) Direct Retry (Per)

Optimisation AspectsHandover Performance (I)

Handover Performance: Inter cell HO - Causes

Percentageof

Quality HO increases

Parameter Optimisation

S1/SFH 1x3 SFH 1x3HAND:HOAVQUAL 5-2 8-2PWRC: LOWTQUAU 3 4

s


Optimisation Aspects Handover Performance (II)

0,0%

10,0%

20,0%

30,0%

40,0%

No Hopping 34,0% 1,1%

S1 / SFH 1x3 32,9% 3,1%

SFH 1x3 30,4% 0,7%

InterCellHO/TCHAssignm IntraCellHO/TCHAssignm

S1/SFH 1x3 SFH 1x3HAND:HOAVQUAL 5-2 8-2PWRC: LOWTQUAU 3 4

Optimisation of Database Parameters Percentage of Intra cell HO decreased again

UL-PC enabled from the beginning (incl. No hopping)

s


Optimisation Aspects Handover Performance (III)

0%

20%

40%

60%

80%

100%

120%

Intra Downlink Quality 54% 62% 74%

Intra Uplink Quality 46% 38% 26%

No Hopping S1 / SFH 1x3 SFH 1x3

S1/SFH 1x3 SFH 1x3HAND: HOAVQUAL 5-2 8-2PWRC: LOWTQUAU 3 4

RXQUAL_DL > 5RXLEV_DL > 35

RXQUAL_UL > 5RXLEV_UL > 31

s


Rate of idle traffic channels per interference band

96.3%

98.7%

98.9%

98.9%

99.3%

1.9%

1.2%

1.1%

1.1%

0.7%

0.9%

0.0%

0.0%

0.0%

0.0%

0.9%

0.0%

0.0%

0.0%

0.0%

94.0% 95.0% 96.0% 97.0% 98.0% 99.0% 100.0% 101.0%

Non Hopping

BFH

SFH

SFH + PC

SFH + PC + DTX

Interference Band 1 Interference Band 2 Interference Band 3 Interference Band 4 Interference Band 5

Optimisations AspectsInfluence of RLC Features on Idle Traffic Channel Measurements

Less percentage of measurements in higher bands Reduction of Interference

s


Implementation of PC & DTX in SFH 1x1

Optimisation AspectsHandover Performance - Influence of PC & DTX

0%

10%

20%

30%

40%

50%

60%

Uplink Quality (Per) 14% 23%

Downlink Quality (Per) 20% 16%

Uplink Strengh (Per) 8% 7%

Downlink Strength (Per) 7% 6%

Distance (Per) 0% 0%

Better Cell (Per) 49% 47%

Direct Retry (Per) 2% 1%

SFH 1x1 PC, DTX SFH 1x1

Increase of no. of uplink quality handovers

Decrease of no. of downlink quality handovers

Power Control was enabled in Downlink additionally

Reduction of Downlink Quality HO

s


Optimisation Aspects Influence of Cell Synchronisation

0,00%

0,20%

0,40%

0,60%

0,80%

1,00%

1,20%

Synchronized cells 1,08 %

Change of HSN 1,07 %

TCH Drop Rate SFH Reuse 1x3

1,00 %

1,09 %

TCH Drop Rate SFH Reuse 1x1

Results of Changing HSN - 1x3 and 1x1 in Reference Cells

Different HSN within one site

No Synchronisation between the cells

Synchronization

No Synchr. No Synchr.

HSN = a HSN = a

HSN = b

s


SFH 1x1: Reduction of frequencies

0,0%

1,0%

2,0%

3,0%

4,0%

5,0%

PC, DTX SFH 1x1 4,7% 1,22%

Carr_Red 3Fr SFH 1x1 4,6% 1,16%

Carr_Red 6Fr SFH 1x1 4,6% 1,21%

Dropped Call Rate TCH Drop Rate

Reduction of 3 Frequencies

Reduction of 6 Frequencies

Optimisation AspectsReduction of Frequencies

s


Optimisation Aspects Drive Test Data - FER Evaluation (I)

FER (full) Samples Percent. Samples

Perc. Sampl Cumul.

Total FER Rate

0% 22883 97,08% 97,08% 0,00%4% 407 1,73% 98,81% 0,07%8% 87 0,37% 99,18% 0,10%

12% 26 0,11% 99,29% 0,11%16% 34 0,14% 99,43% 0,14%20% 46 0,20% 99,63% 0,17%24% 16 0,07% 99,69% 0,19%28% 13 0,06% 99,75% 0,21%32% 10 0,04% 99,79% 0,22%36% 6 0,03% 99,82% 0,23%40% 19 0,08% 99,90% 0,26%44% 6 0,03% 99,92% 0,27%48% 6 0,03% 99,95% 0,28%60% 6 0,03% 99,97% 0,30%88% 6 0,03% 100,00% 0,32%

23571 100% 0,32%

T

T

T

T

T

T

T

T

T

T

T

T

S T

T

T

T

T

T

T

T

T

T

-

T

T

T S

Traffic Channel (TCH)

Slow Associated Control Channel (SACCH)

26 frames = 120 ms

Measurement Mobile reports FER values each SACCH frame

FER can only evaluated in steps of 4 %

s


Optimisation Aspects Drive Test Data - FER Evaluation (II)

Detailed Evaluation of Non Hopping BCCH vs. Hopping TCH :

Hopping Gain can be seen in RxQual- and FER Distribution

• RxQual values of Hopping samples are worse but FER values are better.

• Better FER samples on the hopping carriers.

FER Distribution (RxLev > 10 & RxQual >4)

Non Hopping BCCH vs. Hopping TCH

0,0%

10,0%

20,0%

30,0%

40,0%

50,0%

60,0%

0 4 8 12 16 20 > 20

FER [%]

No Hopping BCCH Hopping TCH 1x3 Hopping TCH 1x1

RxQual Distribution (RxLev >10 & RxQaul >4)

Non Hopping BCCH vs. Hopping TCH

0,0%

10,0%

20,0%

30,0%

40,0%

50,0%

60,0%

70,0%

5 6 7

RxQual

No Hopping BCCH Hopping TCH 1x3 Hopping TCH1x1

s


Optimisation AspectsInfluence of Power Control on RxLev , RxQual Distribution

s


Optimisation AspectsMeasuring FH improvements in the Field

Call drop rates cannot show full FH gains, since SACCH performance is not strongly related to FH

RXQUAL statistics for both uplink and downlink get worse with FH and need to be interpreted -> required RXQUAL

Currently no vendor supports speech quality related FER measurements in the BSS - for downlink, no MS reporting is standardised - for uplink, BS vendor specific implementations are feasible

TEMS drive/walk test can show FH improvement on downlink speech quality

BR6.0 will have measured FER statistics for the uplink and estimated FER statistics for the downlink

s


Summary


s


Summary SFH Planning & Implementation for the Expansion (I)

The benefits of Frequency Hopping have been successfully verified in Field Trials by using different implementation alternatives (BFH, SFH, loose reuse, tight reuse)

Measurements showed improvements of Quality and Capacity (reduction of frequencies).

Quality Feature Capacity Feature

Using FH in real networks provides measures to enhance the reuse (overall reuse including BCCH frequencies of better than 9 maintaining speech quality at the same time)

s


SummarySFH Planning & Implementation for the Expansion (II)

It is recommended to implement BFH as a basic feature.• BFH allows for narrow band combining (e.g. filter combiners) with low insertion loss

useful in noise limited scenarios

It is recommended to use SFH in mature high capacity networks• SFH requires wide band combining (e.g. hybrid combiners)

recommended for interference limited scenarios

Implementation of Features does not replace quality and capacity improvements to be achieved via maintenance and network optimisation activities !

s


SummarySiemens Reference Networks: Synthesiser Frequency Hopping

China

Croatia

Czech Rep.

Germany

Kuwait

RSA

Syria

Taiwan

Thailand

USA

Siemens SFH Networks in

High capacity configurations:

• Network with site configurations up to 4/6/4: 4/4/4, 4/5/4, 4/6/4

• Cells are significantly loaded with traffic

• Call Drop Rate less than 2 % TCH Drop Rate better than 2% (in selected cases better than 1%)

Achievable quality in the networks depends on • coverage situation• available spectrum• Traffic load and traffic distribution• homogeneity of the network and topography of the landscape

s


SummaryAdditional Information

SIEMENS Technical Descriptions Base Station System (TED-BSS BR.xx)

PLMN SBS Radio Network Parameters (SIEMENS ICN Training Institute)

PLMN SBS Performance Measurements (SIEMENS ICN Training Institute)

ETSI GSM Recommendation GSM 05.05, 05.08 and 04.08

30563107 Planning and Optimization of FH in Gsm

Documents

supported

consulting

consulting

frequency

frequency

database parameters

interference

network design