ilsa2 workshop v1

ILSA 2 workshop

1. Plan New sites (1800 MHz)( )2. Retune existing sites (900 MHz)

Copyright 2008 AIRCOM International

START Summaryy

Do analysis existing before Start AFP

AFPDefine reasonable size network (filters)Setup cost matrix based on local knowledge of the area or from previous interference problems. (separation list, cost matrix)Create Best server & Interference Array with y

Low resolution to see overview (for finding problem area)Higher resolution with reasonable size of network and problem to do AFPproblem to do AFP

Run – Check – Run … until get acceptable resultsAFP Can take hours to weeksILSA will not stop until Cost = 0 (only obtain for new networkILSA will not stop until Cost = 0 (only obtain for new network with low capacity and wide bandwidth)

Do Post Anal sis after Finish AFP


Do Post Analysis after Finish AFP

Process ISet Up Frequency BandSetup Prerequisite

Import traffic (if used)Import traffic (if used)Use XML Editor put in column “TRAF”

Import NBR from live or create from NBR WizardC t FiltCreate Filter

PlanRead only

Create Best ServerCreate Interference Table#HO count if used

Check Existing Frequency PlanBest serverWorst connectionWorst connectionWorst interfererTotal interference


Process II

Run ILSA 211. Initialize2. ILSA Menu

Set Plan ListSet Plan ListSet Cost MatrixSet OptionSet OptionRunStop

Continue

▫ Check Frequency Assignment ▫ Apply to Database▫ Check Interference Array & Statisticy▫ OK?▫ NOK?

END


Frequency Assignment Strategiesq y g gAvailable Spectrum for Indosat GSM1800

Ch: 549 – 573 (5 MHz)Ch: 549 573 (5 MHz)Mainly used for Microcells and Indoor sites

Ch: 712 – 786 (15 MHz)Mainly used for the Macrocell layerMainly used for the Macrocell layer

dedicated band for BCCH channels is used in most regions

BCCH1800 755 785BCCH1800 :755-785TCH1800 : 712 - 753

5 BCCHMicrocell

1 guard ch between BCCH &TCH

549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573

4 BCCHIndoor

7 + 9 TCH Chs:Microcell

42 Chs :Macro Layer

BCCH &TCH712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754

Indoor


1 guard ch with other operator

755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786

31 BCCH carriers, Macro Layer

Frequency Band from strategyq y gy


Frequency Bandq y


Array Setting for 1800y g


ILSA WORKSHOP

CHECK EXISTING NETWORKBEFORE AFP NEW SITE 1800

ILSA WORKSHOP

BEFORE AFP NEW SITE 1800


Create Best Server Microsoft Office Excel Worksheet


Worst Connection


Average Connectiong


Worst Interferer


Total Interference


Create NBR for new sites


Create New Site 2G GSM1800


Create New sites GSM 1800 –No FreqNo Freq


Create New sites GSM 1800 –when BCCH Allocatedwhen BCCH Allocated


Select returnable area (1)( )Use Polygon selection


Select returnable area (2)( )


Crate Filter from Selection Expertp


Filters (new sites + read only)( y)


Run Prediction for 2 new sites


Create New Best Server Coverage+ 2 new sites+ 2 new sites


New Service Area


Create Interference table (1)( )


Create Interference table (2) : Select View( )


Create Interference table (3) : Select Filter( )


Create Interference table (4)Select parametersSelect parameters


Create Interference table (5)Select parametersSelect parameters


Save to File


Adding 6 frequencies without retuning surrounding sites

1ST AFP TRIALADD 2 SITES NO RETUNE

Adding 6 frequencies without retuning surrounding sites

ADD 2 SITES NO RETUNE


RUN ILSA2: Initialisation

Database from Filter or XML

One of theseO e o t ese1. Interference Array in Memory2. .ait file (prediction)3. .msmt file (measurement)

#HO occured


Network: FilterWhen click import filter …


Load Interference Table

One of theseO e o t ese1. Interference Array in Memory2. .ait file (prediction)3. .msmt file (measurement)


ILSA MENU: Edit plan listp


First Try to Allocation only for New Sitey y


Cost Matrix Setting : Carrier Costsg

If we define frequency group already and allow all frequencies to beand allow all frequencies to be allocated, this should be blank.


Cost Matrix Setting : Carrier Layer Costsg y

We picked up one layer BCCH only!


Cost Matrix Setting : Equipment Costsg q p

In case we have hierarchy or micro, macro to define cost of ,allocation


Cost Matrix Setting : Neighbour Costsg gAvoid Adj-channel in immediate NBRCo is forbidden

Avoid co-channel in 2nd Order NBRAdj is OK


Cost Matrix Setting : Exception Costsg p

If you don’t have exception in cell layer defined, this will be blanklayer defined, this will be blank


Cost Matrix Setting : Filter Priorityg y

The priority of the filter expressed as a p y pvalue. Filters with a higher value have precedence over filters with a lower value.


Cost Matrix Setting : Handover Count & Intermod CostsIntermod Costs


Cost Matrix Setting : Separation Costsg p


Optionsp


Run … Start APF

ILSA will stop only if Cost = 0 after allocated all frequencies required.


p y q q

Result in : Carrier Assignmentsper Layerper Layer


Apply to Databasepp y


See new Frequency from Mapviewq y p


Visual Check Site 1_S1_


AFP for 2 new sites but no retune

CHECK AFTER 1ST TRIALINTERFERENCE ARRAY

AFP for 2 new sites but no retune

INTERFERENCE ARRAY CHECK


Worst Interferer Before


Worst Interferer After Add 2 sites


Comparison : Before vs Afterp

added 6 frequencies for 2 sites without retune any surrounding sites

No degradation (interference) • Improve probably because adding


when add new sitesp p y g

new 2 sites

Statistically comparison in overviewStatistically comparison in overview

BeforeCategory name Category value Covered area(%)

After add 2 sitesCategory name Category value Covered area(%)

very good 12.00 <= x dB 96.78%

good 9.00 <= x < 12.00 dB 98.76%

very good 12.00 <= x dB 96.58%

good 9.00 <= x < 12.00 dB 98.50%

d dmedium 6.00 <= x < 9.00 dB 99.67%

bad 3.00 <= x < 6.00 dB 99.95%

very bad 0.00 <= x < 3.00 dB 100.00%

medium 6.00 <= x < 9.00 dB 99.28%

bad 3.00 <= x < 6.00 dB 99.49%

very bad 0.00 <= x < 3.00 dB 99.54%y y

• Improve probably because adding new 2 sitesImprove probably because adding new 2 sites• add 6 frequencies for 2 sites without retune any surrounding sites


Adding 6 frequencies and retune surrounding

2ND AFP TRIALADD 2 SITES WITH RETUNE

Adding 6 frequencies and retune surrounding

ADD 2 SITES WITH RETUNE


Plan ListAllow surrounding to be returnableAllow surrounding to be returnable


Progress… 1. Rung 1. Run2. Stop3. Check Assignment4. Apply databasepp y5. Create Interference Array6. Visual & Statistical check7. OK? -> Stop8. NOK? -> Continue


Stop & Checkp


Time Consuming & Loop Check(Process II)(Process II)

Run ILSA 211. Initialize2. ILSA Menu

Set Plan ListSet Plan ListSet Cost MatrixSet OptionSet OptionRunStop

Continue

▫ Check Frequency Assignment ▫ Apply to Database▫ Check Interference Array & Statisticy▫ OK?▫ NOK?

END


Retune 900 MHz

3RD AFP TRIALRetune 900 MHz


Analyze existing plan

BCCH 900 + TCH 900INTERFERENCE ANALYSIS

Analyze existing plan

INTERFERENCE ANALYSIS


Frequency Assignment Strategiesq y g g

Available Spectrum for Indosat GSM900Ch: 1 – 50 (10 MHz)

Shared between the Macrocell, Microcells and Indoor sitesFor the broadcast channel (BCCH)( )

required cluster size BCCH channel > required cluster size TCH channelsdedicated band for BCCH channels is used in most regionsdedicated band for BCCH channels is used in most regions

Total Indosat GSM 900 bandwidth (10 MHz = 50 carriers)Common band:

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50

20 BCCH carriers28 TCH carriers 1 guard ch between BCCH &TCH

1 guard ch with other operator


Define Frequency Bandq y


Array Setting 900 (BCCH+TCH)y g ( )


FH, FD, DTX, ,


Create Best Server 900


Create Worst Connection


Create Average Connectiong


Create Worst Interferer See differences !!See differences !!


Total Interferer


Consider only BCCH900 layer, Find Worst Interferer

BCCH 900 ANALYSISConsider only BCCH900 layer, Find Worst Interferer


Array Settingy g


Best Server


Finding Worst Interfererg


Finding Worst Interfererg

Pointing to area and see worst interferer• Retune Manually• AFP


Use AFP ILSA to retune BCCH 900 in area affected by worst interferer

4TH AFP TRIALRETUNE BCCH 900

Use AFP ILSA to retune BCCH 900 in area affected by worst interferer

RETUNE BCCH 900


Trial Area Red – 1st TargetBlue – 2nd Target gBlack – Read only


SteppCreate Filter Retune_Inner

1st Target to be retunedC Fil R OCreate Filter Retune_Outter

2nd Tier to be RetunableCreate Filter Read_only

Only to be considered as interferer or victim but will not be retuned.

Create Best ServerCreate Interference TableRun ILSAInitializeInitializePlan ListSet up Cost MatrixRunRun ….

Check assignment & Result Apply to DatabaseCh k i t & R lt


Check assignment & Result

Filters Red – 1st TargetBlue – 2nd Target gBlack – Read only


Create Interference Table


Run ILSA


Select Filters to load into ILSA


Select Cells Layersy


Select Interference table from file


Plan List


Filter Priorityy

The priority of the filter expressed as a value. Filters with a higher value have precedence over filters with a lower valueprecedence over filters with a lower value.


Separation Costp


Run (1)( )


Run (2)( )


Results (Inner)Results (Inner)After 10 min After 20 min


Results OutterResults OutterAfter 10 min After 20 min


Continue… tuningg


30 min after Run ILSA•Stop•Apply to database•Create new worst interferer again from new frequencies•Create new worst interferer again from new frequencies•Visual check

Interference area is decreased


END Summaryy

Do analysis existing before Start AFP

AFPDefine reasonable size network (filters)Setup cost matrix based on local knowledge of the area or from previous interference problems. (separation list, cost matrix)Create Best server & Interference Array with y

Low resolution to see overview (for finding problem area)Higher resolution with reasonable size of network and problem to do AFPproblem to do AFP

Run – Check – Run … until get acceptable resultsCan take hours to weeksILSA will not stop until Cost = 0 (only obtain for new networkILSA will not stop until Cost = 0 (only obtain for new network with low capacity and wide bandwidth)

Do Post Anal sis after Finish AFP


Do Post Analysis after Finish AFP

Thank you

Learning by doingg y g


Problem definition, Arrays

SUPPORT SLIDESProblem definition, Arrays


PROBLEM DEFINITION It is sensible to consider the size of the problem that we are attempting to solve. Consider a small example

network. Its characteristics are as follows: S = Number of Sites = 10 C = Number of Cells = 30 = 3 cells / site A = Number of Carrier Allocations Required = 60 = 2 carriers / cellF = Number of Available Frequencies = 20

Therefore Clearly checking every possible frequency plan for a reasonable sizeTherefore, Clearly checking every possible frequency plan for a reasonable size network would take a prohibitive period of time, even if the period of time taken to check a plan were only of the order of 1milli-second. This being the case, algorithms need to be developed to focus processing power on high probability


algorithms need to be developed to focus processing power on high probability solutions. These algorithms are generally known as search algorithms.

Best Server Array y

This array displays the signal strength of the best serving cell at each pixel on the Map View. This decision is based on parameters specified in the Site Database window and in the Array Settings dialog boxin the Array Settings dialog box


Interference Arrayy

When creating one of the Interference arrays, ASSET requires a Best Server array in memory. If this is not the case, a Best Server array will be automatically created. However if you later create subsequent InterferenceHowever, if you later create subsequent Interference arrays after making changes to the network, ASSET does not automatically create a fresh Best Server array.not automatically create a fresh Best Server array.


Per Carrier Interference Array y

For all the interference calculations, ASSET generates an i t di t i t l ll d ' i i t f ' Fintermediate internal array called a 'per carrier interference array'. For each pixel in the array, the serving sub-cell is determined, and for each carrier of the serving sub-cell the worst carrier to interference (C/I) (lowest numerical value) and the total C/I is calculated, taking into consideration all co- and adjacent carriers from all interfering sub-cells. The total C/I is determined by summing the interfering signals in watts and then later converting back to dB. The result is an array such that for each pixel, a list is obtained of serving carriers plus thesuch that for each pixel, a list is obtained of serving carriers plus the worst and total C/I for each carrier. You cannot currently visualise this intermediate array, which no longer exists when all the other selected arrays have been createdlonger exists when all the other selected arrays have been created.


Worst Connection Array y

For each pixel, the serving sub-cell is determined, and for each hopping carrier group the average carrier to interference (C/I) is calculated from the corresponding pixel in the 'per carrier interference array' by convertingpixel in the per carrier interference array , by converting total C/I to BER and calculating the mean. The mean Bit Error Rate is converted back to dB and the hoppingError Rate is converted back to dB and the hopping carrier group with the lowest resultant C/I is presented, that is, it corresponds to the worst of the mean connection C/I values. This interference array type was designed for networks using frequency hopping although it also works for nonusing frequency hopping, although it also works for non-hopping networks. In a non-hopping network, the carrier group can be considered to contain just a single carrier in


g oup ca be co s de ed to co ta just a s g e ca ethe above description.

Average Connection Array g y

For each pixel, the serving sub-cell is determined, and for each hopping carrier group the average carrier to interference (C/I) is calculated from the corresponding pixel in the 'per carrier interference array' by convertingpixel in the per carrier interference array by converting total C/I to BER and calculating the mean. The mean BER is converted back to dB and the average value for allBER is converted back to dB and the average value for all hopping carrier groups is presented.

This interference array type was designed for networks using frequency hopping, although it also works for non-hopping networks In a non hopping network the carrierhopping networks. In a non-hopping network, the carrier group can be considered to contain just a single carrier in the above description.


t e abo e desc pt o

Worst Interferer Array y

For each pixel, the carrier with the worst carrier to interference (C/I) is determined from the corresponding total C/I value in the 'per carrier interference array'. The result is the worst C/I and the sub cell from which theresult is the worst C/I and the sub-cell from which the interference originates.


Total Interference Array y

For each pixel, the total carrier to interference (C/I) is calculated by summing the total C/I per carrier. This array is applicable to both fully-loaded frequency hopping and non hopping networks The calculated C/I is NOT merelynon-hopping networks. The calculated C/I is NOT merely as experienced by any individual subscriber, but rather it represents the total of the interference experienced byrepresents the total of the interference experienced by ALL subscribers at each pixel.


Check List

FiltersSelection expertStaticDynamic Filters byDynamic Filters by

Area ParameterParameterDevices

ArrayBest server, Interference

NBRCreate NBRCreate NBR


ilsa2 workshop v1

Documents