sudarshan_ rf optim
TRANSCRIPT
-
8/8/2019 Sudarshan_ RF Optim
1/186
www.gtllimited.com
GTL Limited
Network EngineeringNetwork Engineering
Training on RF Optimisation
GSMJune 2005June 2005
Presented by : Sudarshan IyengarPresented by : Sudarshan Iyengar
-
8/8/2019 Sudarshan_ RF Optim
2/186
AgendaAgenda
A. Understanding RF Network Cycle
B. Basics of RF Design
C. Why do we need optimization
D. !ptimization "tages
#. $hysical and %ardware !ptimization
F. Data&ase parameter optimization' "pecial (ools
-
8/8/2019 Sudarshan_ RF Optim
3/186
-
8/8/2019 Sudarshan_ RF Optim
4/186
-
8/8/2019 Sudarshan_ RF Optim
5/186
…… Spreadsheet Design ...Spreadsheet Design ...
sua!!y done during Initia! "et#or$ %ui!d
• &in$ budget to 'a!'u!ate the number o( sites)
• *a!'u!ations based on
– subs'riber density+
– tra((i' per subs'riber+
– e,pe'ted gro#th in tra((i'+ et')
-
8/8/2019 Sudarshan_ RF Optim
6/186
…… CW Drive Test/ Model Tuning...CW Drive Test/ Model Tuning...
Purpose
•
Mode! Tuning is used to– -''urate!y a!!o'ate the sites)
– To a'hie.e more a''urate resu!ts (rom the
predi'tion/simu!ation too! dep!oyed)
–
Identi(i'ation o( hotspots/spe'ia! 'o.erage reuirementareas)
– Tuned mode! 'an be used as a ben'hmar$ (or (uture
e,pansions)
-
8/8/2019 Sudarshan_ RF Optim
7/186
Mode! Tuning Pro'ess
•
Setup 'onsists o( Test transmitter (or the parti'u!ar band1GSM 00/3400 6 usua!!y 207
• -ntenna 6 Omni/Pane!+ 'ab!es+ a''essories)
• One 'andidate 'hosen to represent ea'h type o( '!utter
area in the net#or$)
• The '!utter types 'ou!d be urban+ suburban+ rura!+ et')
• The test transmitter is setup on a suitab!e roo(top)
• Test (reuen'y 'hosen and transmitted
• 8ri.e test is 'arried out using re'ei.er or T9MS euipment
set to s'an mode)
…… CW Drive Test/ Model Tuning...CW Drive Test/ Model Tuning...
-
8/8/2019 Sudarshan_ RF Optim
8/186
…… CW Drive Test/ Model Tuning...CW Drive Test/ Model Tuning...
Mode! Tuning Pro'ess
•
8ata 'o!!e'ted 6 R,!e. samp!es aggregated o.er 0;50 mbins)
• The R,!e. measurements are pro'essed and input to the
predi'tion too!)
• *!utter o((set and other parameters are 'orre'ted)
• *orre'tions are made to a'hie.e 6 !o#est possib!e Standard8e.iation .a!ues)
Thus #e ha.e a
-
8/8/2019 Sudarshan_ RF Optim
9/186
…… RF PlanningRF Planning
• The inputs re'ei.ed (rom spreadsheet design and mode!
tuning sur.eys+ is used to prepare a
-
8/8/2019 Sudarshan_ RF Optim
10/186
…… RF PlanningRF Planning
• The output o( the >&8 is
-
8/8/2019 Sudarshan_ RF Optim
11/186
…… RF Site Survey/Drive TestingRF Site Survey/Drive Testing
• sing the inputs pro.ided by the nomina! 'e!! p!an+ the RF
team per(orms
– Sur.eys (or ea'h sear'h ring in the net#or$ to identi(y the
suitab!e 'andidates #hi'h 'an be used (or bui!ding the sites)
– *andidates identi(ied are ran$ed on basis o( their RF
suitabi!ity and other parameters su'h as stru'tura! stabi!ity+!ine o( sight '!earan'e1(or T,+ a''essibi!ity+ 'osts+ et')
– 8ri.e testing may be 'arried out in some 'ases+ to assess
the RF suitabi!ity
– On'e suitab!e 'andidate1s is identi(ied))a'uisition begins@@@
-
8/8/2019 Sudarshan_ RF Optim
12/186
…… RF Planning – The REAL Challenge!!!RF Planning – The REAL Challenge!!!
• -'uisition o( idea! 'andidate poses a rea! 'ha!!enge to the
net#or$ design pro'ess)
• More o(ten than not 'andidates #hi'h are !o#er on priority interms o( RF suitabi!ity are the ones #hi'h get a'uired@@
• O(ten due to a'uisition 'onstraints+ sear'h rings need to bemodi(ied and sometimes e.en the nomina! p!an needs to be'hanged)
•
Thus as an end resu!t the net#or$ bui!t is de.iated (rom the one#hi'h #as origina!!y designed in the nomina! p!an)A@A@@@@B@
-
8/8/2019 Sudarshan_ RF Optim
13/186
Frequency PlanningFrequency Planning
• GSM #or$s on a (reuen'y reuse pattern)
• -s the sites get a'uired and the bui!d pro'ess starts+ the RFp!anners prepare a C(reuen'y p!anD (or the net#or$)
•
8i((erent te'hniues a.ai!ab!e (or (reuen'y p!an 6 a Fi,edP!an+ b >opping P!an 6 (urther di.ided into %aseband >oppingand SynthesiEed Freuen'y >opping
• RF P!anners either manua!!y or by the use o( an -FP1-utomati'
Freuen'y P!anner 'reate a (reuen'y p!an (or the net#or$)
-
8/8/2019 Sudarshan_ RF Optim
14/186
Frequency PlanningFrequency Planning
• -n optima! (reuen'y is 'riti'a! to ensure good RF
per(orman'e o( the net#or$)
• Spe'tra! 'ha!!enges
• &imited band a!!o'ation
• Fast gro#th rate o( subs'ribers/ tra((i' gro#th
• Tighter reuse patterns
-
8/8/2019 Sudarshan_ RF Optim
15/186
RF Optimization/Parametric OptimizationRF Optimization/Parametric Optimization
• 8uring the net#or$ bui!d initia! RF optimiEation is done+
to ensure that the sites bui!t are reasonab!y meetingtheir ob?e'ti.es)
• 8uring the net#or$ bui!d phase it is a!so ensured that
optima! parameter settings are done (or a!! sites toensure good per(orman'e)
• 8etai!ed e,p!anation o( the abo.e to (o!!o#@@
-
8/8/2019 Sudarshan_ RF Optim
16/186
Traffic Planning/Expansion PlanningTraffic Planning/Expansion Planning
• T#o stages (or *apa'ity P!anning I Initia! "et#or$ %ui!d II
Future 9,pansion)
3 Initia! *apa'ity P!an
• Spreadsheet design is used)
• The e,pe'ted tra((i' is 'a!'u!ated based on a 'ertain
amount o( tra((i' assigned per subs'riber 6 say 25 m9)• The tota! tra((i' reuirement is tra((i' per subs'riber
tota! no o( subs'ribers)
• "et#or$ 'apa'ity is based on a 'ertain GOS 6 say 2 )
•
9r!ang % tab!e used to 'a!'u!ate the no) o( TR+ hen'e noo( sites)
-
8/8/2019 Sudarshan_ RF Optim
17/186
Traffic Planning/Expansion PlanningTraffic Planning/Expansion Planning
• T#o stages (or *apa'ity P!anning I Initia! "et#or$ %ui!d
II Future 9,pansion)2 Future 9,pansion
• This 'an a!so be done using spreadsheet design
methodo!ogy+ using a (igure o( e,pe'ted tra((i' gro#th)
• -!ternati.e!y TR additions are done on an ad;ho' basisby studying the tra((i' trend on a #ee$!y/month!y basis)
• In 'ases #here no (urther TR addition is pra'ti'ab!e+
'apa'ity sites are added in the e,isting net#or$)
• Separate p!anning is done (or Tra((i' *hanne!s1T*> and
-''ess *hanne!s 1S8**>)
-
8/8/2019 Sudarshan_ RF Optim
18/186
Inbuilding SolutionsInbuilding Solutions
• I%S is reuired in p!a'es #here indoor 'o.eragereuirement is 'riti'a! and the possibi!ity o( pro.iding'o.erage (rom outdoor sites is not pra'ti'ab!e)
• sua!!y imp!emented (or p!a'es !i$e 'orporate o((i'es+hote!s+ hospita!s+ shopping 'omp!e,es+ et')+ #here both'o.erage and 'apa'ity is essentia!)
• I%S imp!ementations may 'onsist o(•Repeaters 6 &o# 'ost so!ution (or 'o.ering a sma!! area
#ith !ess tra((i'
•Mi'ro'e!!s/Ma'ro'e!!s 6 Separate %TS sites #hi'h 'an be
a sing!e 'arrier Cmi'ro'e!!D or a mu!ti 'arrier Cma'ro'e!!D+
imp!emented in p!a'es #here !arger area needs to be'o.ered and has higher tra((i' reuirement)
-
8/8/2019 Sudarshan_ RF Optim
19/186
Inbuilding SolutionsInbuilding Solutions
• I%S imp!ementations usa!!y dep!oy a
-
8/8/2019 Sudarshan_ RF Optim
20/186
BenchmarkingBenchmarking
• %en'hmar$ing is done (or ha.ing a 'omparison o( o#nnet#or$ #ith 'ompetitorDs net#or$ in terms o(
'o.erage/.oi'e ua!ity)
• %en'hmar$ing is a!so done (or 'omparing o#n net#or$Dsper(orman'e against 'ertain set HPIs or pre.ious!ya'hie.ed per(orman'e targets)
• Spe'ia! too!s !i$e .oi'e euipment is a.ai!ab!e (or .oi'eua!ity ben'hmar$ing)
• For 'o.erage/ua!ity ben'hmar$ing 'ou!d be done usingregu!ar dri.e test and post pro'essing too!s !i$e T9MS and89SH*-T
• "et#or$ Operator/O9M .endor usua!!y sub'ontra'ts thisa'ti.ity to a rd party+ in order to deri.e unbiased resu!ts(rom the e,er'ise
-
8/8/2019 Sudarshan_ RF Optim
21/186
• Statisti'a! data (rom ben'hmar$ing 'an be used as a.a!uab!e input to the net#or$ optimiEation pro'ess)
• The data is used to identi(y #ea$ areas in the net#or$+#hi'h he!ps in de.e!oping strategies (or impro.ing thenet#or$ per(orman'e)
BenchmarkingBenchmarking
-
8/8/2019 Sudarshan_ RF Optim
22/186
-
8/8/2019 Sudarshan_ RF Optim
23/186
Mobi!e *ommuni'ations propagation is impa'ted by :
Path &oss
Re(!e'tion
8i((ra'tion
Mobile Communications PropagationMobile Communications Propagation
PathLossPathLoss
-
8/8/2019 Sudarshan_ RF Optim
24/186
The basi' path !oss is the transmission !oss in (ree spa'e)
Lfsl = 32.4 + 20 log d(in Kms)+ 20logf(in Mhz)
At 900 Mhz, at a distance of !m , Loss = 9." d#
Act$al %&ediction of loss cannot #e done on this, since in a mo#ileen'i&onment the mo#ile ill &ecei'e signals f&om se'e&al &eflections.he a#o'e fo&m$la is onl* 'alid $nde& di&ect Land no &eflection conditions.
d
Path LossPath Loss
ReflectionReflection
-
8/8/2019 Sudarshan_ RF Optim
25/186
-eflection occ$&s hen a %&o%agating elect&omagnetic a'e im%inges $%on as$&face hich has 'e&* la&ge dimensions as com%a&ed to the a'elength of the
%&o%agating a'e. -eflections occ$& f&om s$&face of ea&th, #$ildings,alls and ate&.
he a'e is %a&tiall* a#soed and %a&tiall* &eflected.
Amo$nt of a#so&%tion ill de%end on the &eflection coefficient of the&eflecting s$&face.
-eflection coefficient is f$nction of the mate&ial %&o%e&ties and de%ends ona'e %ola&ization , angle of incidence and the f&e$enc*.
ReflectionReflection
ReflectionReflection
-
8/8/2019 Sudarshan_ RF Optim
26/186
Path !oss (or 2 ;ray Mode! 1 o.er (!at 'ondu'ti.e sur(a'e
ht hr
d
L2&a* = 40 log d / ( 20 log ht + 20 log h& )
Anal*tical fo&m$la, onl* 'alid fo& la&ge& distances ( 0 Km)
Loss inc&eases at la&ge& distance at a &ate of 40d# 1dec.-t 00 MhE+ 30+000m distan'e + ht 300m+ hr 3)5m&(s 333)5 db #hereas &2ray 33K)5 db
his indicates that in 2 &a* %ath , additional loss of " d#.
ReflectionReflection
ReflectionReflection
-
8/8/2019 Sudarshan_ RF Optim
27/186
Re(!e'tion in a'tua! mobi!e en.ironment + #ou!d be (rom mu!tip!e paths) So+ re(!e'tion in mobi!e 'ommuni'ations is Mu!tipath re(!e'tion) RS& #i!! be resu!tant o( !e.e!s 'oming (rom a!! paths)
ReflectionReflection
-
8/8/2019 Sudarshan_ RF Optim
28/186
8i((ra'tion a!!o#s radio signa!s to propagate around the 'ur.edsur(a'e o( earth and behind obstru'tion)
ht hr Shadow region
RS& drops as the re'ei.er mo.es deep into the shado# region
Huygen's principle on phenomenon of diffraction
All %oints on a a'e/f&ont can #e conside&ed as %oint so$&ces fo& the %&od$ction ofseconda&* a'elets, and that these a'elets com#ine to %&od$ce a ne a'e/f&ont inthe di&ection of %&o%agation.
8i((ra'tion is 'aused by the propagation o( se'ondary #a.e!etsinto the shado#ed region
DiffractionDiffraction
-
8/8/2019 Sudarshan_ RF Optim
29/186
>i!!s+ Mountains+ %ui!dings #i!! 'ause $ni(e edge di((ra'tion In a Mobi!e en.ironment most o( the di((ra'tion is $ni(e edge)
Diffraction is of two types in general
Smooth Sphere 8i((ra'tion Hni(e 9dge 8i((ra'tion
Smooth Sphere 8i((ra'tion
8i((ra'tion ta$es p!a'e through a!most a (!at sur(a'e)
Hni(e 9dge 8i((ra'tion
DiffractionDiffraction
CalculationofDiffractionLossCalculationofDiffractionLoss
-
8/8/2019 Sudarshan_ RF Optim
30/186
Fresne! Eone geometry
-rea around the &OS #ithin #hi'h a di((ra'tion 'an resu!t into
antiphase1340 deg 'ondition is the (irst (resne! Eone)
I( an ob?e't is #ithin the (resne! Eone or 'omp!ete!y b!o'$s the Eone+ then a!so energy#i!! arri.e at the re'ei.er but #i!! di((ra'tion !oss)
In Mobi!e en.ironment+ #e are not #orried about '!earan'e+ but on!y #ith the !oss)
ht hr
Calculation of Diffraction LossCalculation of Diffraction Loss
CalculationofDiffractionLossCalculationofDiffractionLoss
-
8/8/2019 Sudarshan_ RF Optim
31/186
h
d1 d2
&esnel diff&action %a&amete& (')
Indi'ates the position o( the ob?e't #ith re(eren'e to the (resne!
Eones 1 0 means + ob?e't tip on &OS+ 3 means tip on 3st (resne!Eone on upper side)
2 ( d1 + d2 )
d1.d2hv =
&om ' , e can com%$te the diff&action loss.
( all values in "m" )
Calculation of Diffraction LossCalculation of Diffraction Loss
CalculationofDiffractionLossCalculationofDiffractionLoss
-
8/8/2019 Sudarshan_ RF Optim
32/186
Re!ation o( L.L #ith di((ra'tion !oss 1 graphi'a! )
K n i f e e d g e d i f f r a c t i o n g a i n ( G
a d B )
Fresnel diffraction parameter v
-3 -2 -1 0 1 2 3 !-30
-2!
-20
-1!
-10
-!
0
!
Calculation of Diffraction LossCalculation of Diffraction Loss
-
8/8/2019 Sudarshan_ RF Optim
33/186
-adio a'e hen im%inges on a &o$gh s$&face , &eflected ene&g*is s%&ead o$t in all di&ections d$e to scatte&ing.
his is the &eason act$al -L in a mo#ile en'i&onment is oftenst&onge& then hat is %&edicted #* &eflection and diff&action.
#ects s$ch as Lam% %osts and t&ees tend to scatte& ene&g* in
all di&ections, the&e#* %&o'iding additional &adio ene&g* at the&ecei'e&.
ScatteringScattering
-
8/8/2019 Sudarshan_ RF Optim
34/186
o&m$la5s desc&i#ed ea&lie& a&e #ased on sim%le models of the &adio %ath.
o&m$la5s don5t ta!e ca&e of the t*%e of the te&&ain of the &adio %ath.
-ealistic method of %&ediction o$ld #e to $se em%i&ical data of &adioa'e %&o%agation o'e& 'a&io$s t*%es of te&&ain and land $sage.
6m%i&ical data of this t*%e e&e collected #* !$m$&a f&om hiscom%&ehensi'e &adio a'e %&o%agation meas$&ements
Path Loss PredictionPath Loss Prediction
OkumuraModelOkumuraModel
-
8/8/2019 Sudarshan_ RF Optim
35/186
!$m$&a de'elo%ed a set of c$&'es gi'ing the atten$ation in e7cess toL in an $an a&ea ith #ase station effecti'e height of 200m and and
mo#ile antenna height of 3m.
hese c$&'es gi'e the loss as a f$nction of f&e$enc* and distance f&om#ase station)
Okumura ModelOkumura Model
OkumuraModelOkumuraModel
-
8/8/2019 Sudarshan_ RF Optim
36/186
Okumura ModelOkumura Model
Ok M dlOkumuraModel
-
8/8/2019 Sudarshan_ RF Optim
37/186
Path !oss (or di((erent heights 'an be 'a!'u!ated by these 'ur.es byusing the (ormu!a)
8ath Loss = Lfsl + A(f,d) / (hte) / (h&e)
G1hte and G1hre are the e((e'ti.e %ase Station and MS antenna heights
"#hte) $ 20 log 1000m % hte % 10mhte2
( )
"#hre) $ 10 log hre & 3mhr e !
( )
hr e
!
( )"#hre) $ 20 log hre & 3m
Okumura ModelOkumura Model
OkumuraModelOkumuraModel
-
8/8/2019 Sudarshan_ RF Optim
38/186
:hat is 6ffecti'e Antenna ;eight <
3 m 1! m
hmsl
(erage ground leel #hag)
hte
hte $ (ntenna height a*oe msl#hmsl) - aerage ground leel #hag)
( average ground level is calculated ithin ! # 1$%m )
Okumura ModelOkumura Model
OkumuraModelOkumura Model
-
8/8/2019 Sudarshan_ RF Optim
39/186
O$umura 'ur.es are on!y app!i'ab!e (or urban areas)
For other terrains+ O$umura has pro.ided 'orre'tion (a'tors) The 'orre'tion (a'tors are pro.ided (or types o( terrain in the
(orm o( 'ur.es re!ated to (reuen'y)
Open -rea : 'orresponds to a rura! + desert $ind o( terrain
uasi Open -rea : 'orresponds to rura! + 'ountryside $ind o( terrain
Suburban
Okumura Model
OkumuraModelOkumura Model
-
8/8/2019 Sudarshan_ RF Optim
40/186
8ath Loss(o,,s) = Lfsl + A(f,d) / (hte) / (h&e) / (a&ea)
8ath loss fo& othe& te&&ain5s
Okumura Model
Okumura ModelOkumura Model
-
8/8/2019 Sudarshan_ RF Optim
41/186
+onclusion
im%lest, #est and acc$&ate %&ediction model #$t onl* fo& s%ecificte&&ain5s.
lo &es%onse to &a%id changes in te&&ain.
Model is fai&l* good fo& $an and s$#$an a&eas, #$t not as good in&$&al a&eas.
tanda&d de'iations #eteen %&edicted and meas$&ed loss 'al$es 0d to 4 d.
Model is mo&e g&a%hical than mathematical, fo& com%$tation e needfo&m$la5s not g&a%hs.
HataModelHataModel
-
8/8/2019 Sudarshan_ RF Optim
42/186
;ata model is an em%i&ical fo&m$lation of the g&a%hical %ath loss data%&o'ided #* !$m$&a.
;ata %&esented the $an a&ea %&o%agation loss as a standa&d fo&m$laand s$%%lied co&&ection e$ations fo& a%%lications fo& othe& sit$ations.
o&m$la5s a&e designed fo& com%$te& $sage, #$t the* a&e onl* &o$gh
a%%&o7imations of !$m$&a5s c$&'es.
ince e&&ain t*%es %&ofiles a&e %&acticall* infinite, modeling of thetool $sed fo& %&ediction #eca$se essential #$t ta!ing se'e&almeas$&ements and se'e&al times.
Hata ModelHata Model
Hata ModelHata Model
-
8/8/2019 Sudarshan_ RF Optim
43/186
rban terrain
L($an) = >9."" + 2>.> log fc / 3.?2 log hte / a(h&e) + ( 44.9 / >."" log hte ) log d
fc = fre&uenc' in * ( 1$ # 1 *)
hte = B, antenna height ranging !m to 2m
hre = effective receiver antenna height ranging 1m to 1m
d = ransmitter receiver separation distance (1 # 2 %m )
a(hre ) = correction factor for effective mo-ile antenna height hich is a function of si*e of the coverage area in d-
Small to medium city
a(hre) = (1.1 log fc # . ) hre # ( 1.$/ log fc # .0 ) d-
For large city
a(hre) = !.2 ( log 11.$ hre ) # .2
Hata ModelHata Model
-
8/8/2019 Sudarshan_ RF Optim
44/186
*orre'tion (or Suburban N Rura! terrains
&oss (or Rura! Open -rea
&oss (or S%R%-"
3(su-) = 3 (ur-an) # 2 4 log (fc520) 6 # $.2
3(ro) = 3 (ur-an) # .0 ( log fc) # 10.!!log fc # .2
&oss (or Rura! uasi;Open -rea
3(r&o) = 3 (ur-an) # .0 ( log fc) # 10.!!log fc # !$.2
COST231 HataModelCOST231-HataModel
-
8/8/2019 Sudarshan_ RF Optim
45/186
L($) = 4>.3 + 33.9 log fc / 3.?2 log hte / a(h&e) + ( 44.9 / >."" log hte ) log d +@m
fc = f&e$enc* in M;z ( "00 / 2000 M;z) hte = antenna height &anging 30m to 200m
h&e = effecti'e &ecei'e& antenna height &anging m to 0m d = &ansmitte& &ecei'e& se%a&ation distance ( / 20 !m )a(h&e ) = co&&ection facto& fo& effecti'e mo#ile antenna height hich
is a f$nction of size of the co'e&age a&ea in d#@m = @o&&ection facto& fo& cit* size
a(h&e) = (. log fc / 0. ) h&e / ( ."> log fc / 0.? ) d#@m = 0 d# fo& medi$m cit* and s$#$an cente&s ith mode&ate t&ee densit@m = 3 d# fo& met&o%olitan cente&s.
For rura! areas + the ear!ier (ormu!as #i!! app!y
@ /23o&!ing committee de'elo%ed an e7tended 'e&sion of;AA model fo& f&e$encies $% to 2 ;z.
COST 231 - Hata ModelCOST 231 - Hata Model
Path Loss Predictions for GSMPath Loss Predictions for GSM
-
8/8/2019 Sudarshan_ RF Optim
46/186
election of models fo& %&edicting %ath loss fo& M ill de%end on thecell &anges.
M has 3 cell &anges and diffe&ent %&ediction model fo& each
La&ge @ells
mall @ells
Mic&ocells
Lrg CllLargeCells
-
8/8/2019 Sudarshan_ RF Optim
47/186
Antenna is installed a#o'e the ma7im$m height of the s$&&o$nding &oof
to%s. 8&o%agation is mainl* #* diff&action and scatte&ing at &oof to%s in the'icinit* of the mo#ile i.e. the main &a*s %&o%agate a#o'e the &oof to%s.
@ell &adi$s is mainl* !m and no&mall* e7ceeds 3 !m. ;ata5s model and the @ 23/;ata model can #e $sed to calc$late
%ath loss in s$ch cells.
Large CellsLarge Cells
Small CellsSmall Cells
-
8/8/2019 Sudarshan_ RF Optim
48/186
Antenna is sited a#o'e the median #$t #elo the ma7im$m height ofthe s$&&o$nding &oof to%s.
8&o%agation mechanism is same as la&ge cell Ma7im$m &ange is t*%icall* less than / 3 !ms. ;ata model cannot #e $sed since it is a%%lica#le a#o'e !m. @ 23/:alfish/B!egami model is $sed fo& &adi$s less than "!ms in
$an en'i&onment.
COST 231 - Walfish-Ikegami ModelCOST 231 - Walfish-Ikegami Model
-
8/8/2019 Sudarshan_ RF Optim
49/186
7ithout free 38, -eteen B, and ,
&e$enc* (f) = ?00 / 2000 M;z&ansmitte& height (hte) = 4 / "0mMo#ile height (h&e) = / 3mCistance (d) = 0.02 / " !m;eight of #$ildings ;&oof (m)
:idth of &oad (m)$ilding se%a&ation # (m)-oad o&ientation ith &es%ect to the di&ect &adio %ath 8hi (o)&ee %ace Loss (Lfsl)
8ath Loss = Lfsl + L&st + Lmsd
L&ts D &oof/to%/to/st&eet diff&action and scatte& lossLmsd D m$ltisc&een diff&action loss
COST 231 - Walfish-Ikegami ModelCOST 231 - Walfish-Ikegami Model
-
8/8/2019 Sudarshan_ RF Optim
50/186
:ith a f&ee L #eteen #s and ms ( t&eet @an*on )
8ath Loss = 42.> + 20 log(d) + 20 log(f) fo& d = 0.020 !m
Indoor LossIndoor Loss
-
8/8/2019 Sudarshan_ RF Optim
51/186
Additional loss hich occ$&s at 900 M;z hen mo'ing into a ho$se on the#ottom floo& on ."m height f&om the st&eet.
Bndoo& loss nea& indos ( E m ) is t*%icall* 2 d#.
$ilding loss as meas$&ed #* inish 8 'a&ies #eteen 3 d# and /?d#ith an a'e&age of ?d# ta!en o'e& all floo&s and #$ildings.
Bn o$& %&edictions and calc$lations, as %e& M &ecommendations e illconside& "d# as an a'e&age indoo& loss.
MicrocellsMicrocells
-
8/8/2019 Sudarshan_ RF Optim
52/186
@ell in hich the #ase station antenna is mo$nted gene&all* #elo &oofto% le'el.
8&o%agation is dete&mined #* diff&action and scatte&ing a&o$nd#$ildings ie. the main &a*s %&o%agate in st&eet can*ons.
Mic&ocells ha'e a &adi$s in the &egion of 200 / 300m . Mic&ocells can #e s$%%o&ted #* smalle& and chea%e& 5s.
MicrocellsModelMicrocellsModel
-
8/8/2019 Sudarshan_ RF Optim
53/186
7ith a free 38, -eteen -s and ms ( ,treet 9an'on )
8ath Loss (M 900 ) = 0. + 20 log(d) fo& d = 0.020 !m8ath Loss (C@ ?00 ) = 0. + 20 log(d) fo& d = 0.020 !m
8&o%agation loss in mic&ocells inc&eases sha&%l* as the &ecei'e& mo'eso$t of L , (e7 D a&o$nd a st&eet co&ne& ).
20d# of loss co$ld #e added %e& st&eet co&ne&, $% to to o& th&ee co&ne&s.
e*ond, this the @23 / :alfish B!egami model sho$ld #e $sed
Microcells ModelMicrocells Model
FadingFading
-
8/8/2019 Sudarshan_ RF Optim
54/186
A mo#ile &adio signal en'elo%e has contin$os 'a&iations.hese 'a&iations contin$o$sl* fl$ct$ate the signal le'el and is &efe&&ed
to as the fading %henomenon.
ading in mo#ile en'i&onment is of 2 t*%esD
Sma!! S'a!e Fading
&og;norma! Fading
-
8/8/2019 Sudarshan_ RF Optim
55/186
Small Scale FadingSmall Scale Fading
-
8/8/2019 Sudarshan_ RF Optim
56/186
,*sering the Short-term fading with reference to aerage leel
Ray!eigh distribution
r o
r o #d*) $ (erage - instantaneous fluctuations
#small-scale fading)
&o &anges in 40 d# ( 0 d# a#o'e and 30 d# #elo the a'e&age )
&o follos a -a*leigh/dist&i#$tion , since gene&all* signals a&&i'e f&om&eflections onl*, hence small/scale fading is often called -a*leigh/2ading.
Small Scale FadingSmall Scale Fading
-
8/8/2019 Sudarshan_ RF Optim
57/186
Ray!eigh distributionAs %e& -a*leigh dist&i#$tion inc&ease in fade de%th is in'e&sel*
%&o%o&tional to the %&o#a#ilit* (e7D 0 d# fade ma* occ$& fo& 40 Fof the time, he&e the %&o#a#ilit* of 40d# fade o$ld #e 0 F )
avg level
min recv level
of rcvr deepest fades ( t'picall' ! d- )
:rea of poor &ualit'
;r
8&o#a#ilit* that fade de%ths ill ente& a&ea of %oo& $alit* is&e$i&ed to #e less than 0 F.
Fade argin
Small Scale FadingSmall Scale Fading
-
8/8/2019 Sudarshan_ RF Optim
58/186
Bf %&o#a#ilit* of small/scale fades is mo&e #elo the minim$m &e$i&edsignal le'el, then this co$ld &es$lt in disto&ted s%eech.
o ens$&e this %&o#a#ilit* is less than 0 F , &ansmit 8oe& sho$ld #ead$sted acco&dingl* to achie'e a high fade ma&gin.
%ace Ci'e&sit* is $ite effecti'e fo& this !ind of fades.
-a*leigh dist&i#$tion onl* occ$&s hen the&e a&e all &eflected a'es andno di&ect L signal. Bf the&e is a di&ect L signal %&esent ith&eflections, then it is -icean dist&i#$tion of fading hich is less se'e&e ,since the di&ect com%onent is &elati'el* m$ch st&onge& than &eflected
a'es and ill &est&ict dee% fades.
S S gg
Log-normal FadingLog-normal Fading
-
8/8/2019 Sudarshan_ RF Optim
59/186
ime
,,
og-normal ading
mall /scale signal 'a&iation hen a'e&aged o$t is called the local mean and ise7%&essed in log scale of %oe& , and is called Log/no&mal fading.
Log/no&mal 'a&iation is d$e to the te&&ain conto$& #eteen the #s and ms. Bf the te&&ain is an o%en a&ea, then the change in signal ill #e ith distance onl*,
#$t no&mall* the&e a&e o#st&$ctions ( #$ildings, t&ees etc. ) hich ca$se a &a%id'a&iation of signal f&om its local mean o'e& an a&ea of " to "0m
-
8/8/2019 Sudarshan_ RF Optim
60/186
-
8/8/2019 Sudarshan_ RF Optim
61/186
Reducing Time Dispersion IssuesReducing Time Dispersion Issues
-
8/8/2019 Sudarshan_ RF Optim
62/186
%timization and @o$nte&meas$&es fo& ime Cis%e&sion is something'e&* inte&esting and can &es$lt into se'e&al iss$es li!e disto&ted 'oice,
echo and e'en d&o%%ed calls GG
@e&tain co$nte&meas$&es hen ado%ted in the %lanning stage can &ed$ceo& eliminate these iss$es.
Bf %&o#lems occ$& late& on, then o%timization needs to #e done. &o$#leshooting m$lti%ath %&o#lems is a #ig iss$e in li'e neto&!s.
Reducing Time Dispersion IssuesReducing Time Dispersion Issues
-
8/8/2019 Sudarshan_ RF Optim
63/186
Site &o'ation
Bdentif* %otential &eflecto&s in the %&edicted cell a&ea.
Locate sites fo& nea& &eflecto&s, this ill #&ing the &eflectionsithin the indo.
Reducing Time Dispersion IssuesReducing Time Dispersion Issues
-
8/8/2019 Sudarshan_ RF Optim
64/186
8ire'tiona! -ntennas1Se'toriEation
Hsing ecto&ed cells config, ith the di&ectional antenna %ointing aa* f&om the &eflecto&.
Antenna5s f&ont/#ac! &atio is a 'e&* c&itical %a&amete&.
Reducing Time Dispersion IssuesReducing Time Dispersion Issues
-
8/8/2019 Sudarshan_ RF Optim
65/186
O.er 7ater %odies ime dis%e&sion o'e& ate& can ma!e the $alit* o&st
@o'e&ing the a&ea f&om the othe& side of the ate& #od* ill a'oidla&ge %ath diffe&ences #eteen &eflected signals.
ide lo#es can still &es$lt into %&o#lems, he&e hando'e&s sho$ldta!e ca&e off, #* %&o%e&l* setting neigh#o&sI %a&amete&s
Reducing Time Dispersion IssuesReducing Time Dispersion Issues
-
8/8/2019 Sudarshan_ RF Optim
66/186
Ti!ting -ntennas
ilting Antenna ill &ed$ce ene&g* &adiated toa&ds the &eflecto&. Antenna5s can #e tilted ho&izontall* o& 'e&ticall*.
;o&izontal tilt ill &ed$ce the co'e&age to a la&ge e7tent, hence'e&tical tilt is the most %&efe&&ed one.
Redu'ing Output Po#er
-ed$ction in o$t%$t %oe& ill &ed$ce the ene&g* f&om #oth di&ect as ell &eflected signal. ;ence, 81M ill not change.
Doppler ShiftDoppler Shift
-
8/8/2019 Sudarshan_ RF Optim
67/186
he shift in f&e$enc* &elati'e to the s%eed of the mo#ile %hone isCo%%le& hift.
fd = v
fd = ,hift in fre&uenc' in *
v = speed of the mo-ile in m5s = avelength in m
Act$al &ecei'ed ca&&ie& f&e$enc* = fc + fd, hen mo#ile is mo'ingtoa&ds the t&ansmitte&.
Act$al &ecei'ed ca&&ie& f&e$enc* = fc / fd, hen mo#ile is mo'ingaa* f&om the t&ansmitte&.
he&e is no shift , hen the 'ehicle is mo'ing %e&%endic$la& to theangle of a&&i'al of the t&ansmitted signal.
Frequency PlanningFrequency Planning
-
8/8/2019 Sudarshan_ RF Optim
68/186
Ob?e'ti.e
%tim$m $ses of -eso$&ces
-ed$ce Bnte&fe&ence
Frequency PlanningFrequency Planning
-
8/8/2019 Sudarshan_ RF Optim
69/186
F=1
F=2
F=!
F=
Co - Channel Re-use factorCo - Channel Re-use factor
-
8/8/2019 Sudarshan_ RF Optim
70/186
:
:
> = ?
+ $ d*
> = ,&rt ( ! @ A )
Adjacent-Channel Re-use CriteriaAdjacent-Channel Re-use Criteria
-
8/8/2019 Sudarshan_ RF Optim
71/186
Adacent A-@J5s sho$ld not #e $sed in the same cell
Bt ill ha'e no %&o#lems in Conlin!, #$t ill ha'e high &is! of $%lin!inte&fe&ence (d$e to mandato&* $%lin! %oe& cont&ol ).
Bf Conlin! d*namic %oe& cont&ol is not $sed
- 40 d*m # +a $ 20 )
- 0 d*m # +a $ -20 )
! d*m
33 d*m
ince all the A-@J5s in a cell a&ef&ame s*nched, imeslot n$m#e&sill align on all the A-@n5s
jj
Adjacent-Channel Re-use CriteriaAdjacent-Channel Re-use Criteria
-
8/8/2019 Sudarshan_ RF Optim
72/186
Adacent A-@J5s can #e $sed in adacent cells, #$t as fa& as %ossi#lesho$ld #e a'oided.
As s$ch se%a&ation of 200 Khz is s$fficient, #$t ta!ing intoconside&ation the %&o%agation effects, as facto& of %&otection >00
Khz sho$ld #e $sed.
Bn the o&st, Adacent A-@J5s can also #e $sed in adacent cells #*setting a%%&o%&iate hando'e& %a&amete&s ( disc$ssed late& ino%timization)
8&acticall* not %ossi#le in most of the neto&!s d$e to tight &e$se
Cell ConfigurationCell Configuration
-
8/8/2019 Sudarshan_ RF Optim
73/186
mnidi&ectional @ell
B,
ecto&ial @ell
B,
&o# gain -ntennas &esser penetration/dire'ti.ity Re'ei.es Int (rom a!! dire'tions &o#er imp!ementation 'ost
>igh gain -ntennas >igher penetration/dire'ti.ity Re'ei.es Int (rom !esser dire'tions >igher imp!ementation 'ost
-
8/8/2019 Sudarshan_ RF Optim
74/186
Sectored CellsSectored Cells
-
8/8/2019 Sudarshan_ RF Optim
75/186
:1
:2
:!!
/
B1
B2
B! !
/
9
1
92
9!!
/
Re'ei.esInter(eren'e(rom !esserdire'tions)
Re-use PatternsRe-use Patterns
-
8/8/2019 Sudarshan_ RF Optim
76/186
-e/$se 8atte&ns ens$&es the o%tim$m se%a&ation #eteen @o/@hannels.
-e/$se %atte&n is a fo&mation of a cl$ste& ith a %atte&n of f&e$enc*dist&i#$tion in each cell of the cl$ste&.
ame cl$ste& %atte&n is then &e/$sed.
Pre(erred Re;use Patterns
mni / @ells D 3 cell, cell, 2 cell, 4 cell, 9 cells etc
ecto& / @ells D 319 , 412, 12
3/9 Re-use Pattern3/9 Re-use Pattern
-
8/8/2019 Sudarshan_ RF Optim
77/186
(1
(2(3 51
53+1
+2+3
(1
(2(3 51
5253+1
+2+3
(1
(2(3 51
5253+1
+2+3(1
(2(3 51
5253+1
+2+3 (1
51
5253
(1
(2(3
52 +1
+2+3
+2+3+2+3 +2+3
(1
Exercise !!!Exercise !!!
-
8/8/2019 Sudarshan_ RF Optim
78/186
(1
(2(3 51
53+1
+2+3
(1
(2(3 51
5253+1
+2+3
(1
(2(3 51
5253+1
+2+3(1
(2(3 51
5253+1
+2+3 (1
51
5253
(1
(2(3
52+1
+2+3
+2+3+2+3 +2+3
(1
Hsing A-@J5s to9 , do the channel allocation fo& the #elo cells$sing 319 %atte&n
Frequency Allocation in 3/9 patternsFrequency Allocation in 3/9 patterns
-
8/8/2019 Sudarshan_ RF Optim
79/186
Adacent @hannel Bnte&fe&ence is 'e&* diffic$lt to a'oid ithinthe cl$ste& itself.
1
3
2
6!
4
7
4/12 Reuse Patterns4/12 Reuse Patterns
-
8/8/2019 Sudarshan_ RF Optim
80/186
D1
D3
51
53
+1
+2+3 D1
(1
(2(3 51
5253+1
+2+351
5253 (1
(2(3+1
+2+3 +1
D1
D2D3
D2D35253 5253
D2 +1
+3
52
D2D3(1
(2(3 51
5253
+2 D1
D2D3(1
ExerciseExercise
-
8/8/2019 Sudarshan_ RF Optim
81/186
Hsing A-@J5s > to2 do the channel allocation fo& the #elo cells$sing 412 %atte&n.
D1
D2D3 +1
+351
5253
+1
+2+3 D1
D2D3(1
(2(3
(1
(2(3 51
5253+1
+2+351
5253 (1
(2(3+1
+2+3 +1
D1
D2D3
51
5253
+2D1
D2D3
D2D35253 5253
(1
4/12 Pattern Channel Allocation4/12 Pattern Channel Allocation
-
8/8/2019 Sudarshan_ RF Optim
82/186
1
3!
2 7
4
1112
10 6
412 %atte&n a'oids adacent channels in adacent cells
Reuse Patterns ConclusionReuse Patterns Conclusion
-
8/8/2019 Sudarshan_ RF Optim
83/186
&arger reuse patterns gi.e redu'tion in inter(eren'e
Re;use patterns be'omes more e((e'ti.e #ith se'toria! 'e!!'on(igurations)
To imp!ement !arge patterns 1 !i$e /32+ Q/23 + more 'hanne!sare reuired)
So #ith !ess resour'es+ the best #ay to p!an is :
3) se optimum no o( 'hanne!s per 'e!!)2) Thus+ in'rease the pattern siEe)
Critical Factors for good RF NetworkCritical Factors for good RF Network
-
8/8/2019 Sudarshan_ RF Optim
84/186
• Grid based RF design)
• Maintain standard aEimuths #hi!e se'toriEing 'e!!s 6 This
ma$es (reuen'y p!an easier• *orre't 'hoi'e o( antenna type (or spe'i(i' 'o.erage
reuirements)
• se o( optima! antenna heights 6 Shou!d be su((i'ient to
'ater to the 'o.erage area+ but shou!d not e,'eed thereuirement+ e!se it resu!ts into !arge spi!!o.ers andinter(eren'e+ ma$ing reuse di((i'u!t@@
• se optima! ti!t 6 9!e'tri'a! ti!t as (ar as possib!e) Insome 'ases 'ombination o( e!e'tri'a! and me'hani'a! ti!ts
-
8/8/2019 Sudarshan_ RF Optim
85/186
Quality of ServiceQuality of Service
-
8/8/2019 Sudarshan_ RF Optim
86/186
8ffect of 9,S :
Cissatisfied @$stome&s### @$stome&s face desc&i#es *o$& %&ofit
c$&'e/// Cissatisfied c$stome& %&e'ents 0 ne
-e'en$e/// @$stome& itcho'e&s/// Less Je @$stome&s/// @ost of C&o%%ed @alls/// @ost of loc!ed @alls
Importance of RF OptimizationImportance of RF Optimization
-
8/8/2019 Sudarshan_ RF Optim
87/186
• - %timization is a contin$o$s and ite&ati'e %&ocess.
• Main oal o achie'e %e&fo&mance le'els to a ce&tain set
standa&d.• Jeto&! s$#sc&i#e&s e7%ect i&eline1nea& i&eline $alit*.
• Jeto&! s$#sc&i#e&s also e7%ect 00 F a'aila#ilit* at all gi'entimes.
• - neto&! o%timization is a %&ocess to t&* and meet thee7%ectation of s$#sc&i#e&s in te&ms of co'e&age, o,neto&! a'aila#ilit*.
• - o%timization also aims to ma7imize the $tilit* of thea'aila#le neto&! &eso$&ces.
• 6ach o%e&ato& has a ce&tain set of decided K8Bs (Ke*8e&fo&mance Bndicato&s) #ased on hich the o%e&ato& g$agesthe %e&fo&mance of his neto&!.
Importance of RF OptimizationImportance of RF Optimization
-
8/8/2019 Sudarshan_ RF Optim
88/186
• -1Access Jeto&! K8Bs can #e #&oadl* classified intoth&ee t*%es
a) Access &elated K8B#) &affic1-eso$&ce Hsage &elated K8B
c) ;ando'e& &elated K8B
• 67am%les of access K8B
a)C@@; C&o% &ate #) @all set$% s$ccess &atec)C@@; loc!ing, etc.
• 67am%les of &affic K8B
a)@; C&o% -ate #) @all s$ccess &ate
c)@; loc!ing, etc.• 67am%les of hando'e& %e&fo&mance K8B
a);ando'e& $ccess &ate #) ;ando'e& fail$&e &ate.
c);ando'e& %e& ca$se, %e& neigh#o$&, etc.
Importance of RF OptimizationImportance of RF Optimization
-
8/8/2019 Sudarshan_ RF Optim
89/186
• A%a&t f&om the K8Bs mentioned ea&lie& the o%e&ato& ma*ha'e his on set of c$stom K8Bs hich the o%e&ato& feels is
c&itical to g$age the %e&fo&mance of his neto&!.• - o%timization %&ocess d&i'es the effo&t to achie'e and
maintain the neto&! %e&fo&mance K8B.
• %timization can #e #&oadl* di'ided into 3 catego&ies, asfollos
a) ;a&da&e %timization
#) 8h*sical %timization
c) Cata#ase18a&amete& %timization
• ene&all* the acti'ities mentioned a#o'e a&e done in %a&allel.
Bn some cases one ma* %&ecede the othe&.
-
8/8/2019 Sudarshan_ RF Optim
90/186
Network Optimization Cycle…Network Optimization Cycle…
-
8/8/2019 Sudarshan_ RF Optim
91/186
!ptimization "tages!ptimization "tages
RF $lanning
Network Rollo+t
'B+ild $hase
Nominal Cell Design
RF Fine t+ning
Data&ase
parameter optimization
$hysical'
%ardware
!ptimization
Network $re 4
!ptimization
(ra0c !ptimization
-
8/8/2019 Sudarshan_ RF Optim
92/186
-
8/8/2019 Sudarshan_ RF Optim
93/186
Hardware Optimization -Hardware Optimization -Typical Hardware Problems
-
8/8/2019 Sudarshan_ RF Optim
94/186
• Bn most cases, ha&da&e fail$&es on a 1@ o& an* %a&tof the access neto&! ala&ms a&e gene&ated at the M@,
hich hel% in identif*ing the fa$lt• Bn some cases, the&e a&e no ala&ms gene&ated
• Ke* statistics f&om M@- co$ld %oint toa&ds ha&da&efail$&es *%ical statistics hich indicate s$ch %&o#lems a&e
a) 8oo& Assignment $ccess1;igh Assignment fail$&e &ate
#) ;igh @;1C - Loss
c) ;igh hando'e& fail$&e &ate
d) Loe& call 'ol$me1t&affic on the cell
-
8/8/2019 Sudarshan_ RF Optim
95/186
Hardware Optimization -Hardware Optimization -Typical Hardware Problems
-
8/8/2019 Sudarshan_ RF Optim
96/186
• 8ath #alance %&o#lems his is also one of the commonca$ses fo& %oo& cell %e&fo&mance.
%ath #alance is %egged as an M@- statistic on a cell #asisene&al fo&m$la is %ath #alance=$%lin! %athloss donlin!%athloss.
Bn Moto&ola %ath#alance= %athloss+0.
he&e %athloss = $%lin! %athloss donlin! %athloss.$%lin! %athloss = act$al Ms 7%oe& &7le'S$l
donlin! %athloss = act$al s 7%oe& &7le'Sdl
Bt is desi&a#le to ha'e the %athloss 'al$e as O0P hich&e%&esents a #alanced %ath. ;oe'e& a de'iation of +1/ 0 isacce%ta#le
Hardware Optimization -Hardware Optimization -Typical Hardware Problems
-
8/8/2019 Sudarshan_ RF Optim
97/186
• 8ath #alance %&o#lems Bf the %ath#alance is #elo 00 o&a#o'e 20, it indicates that the&e co$ld #e a %&o#lem in
eithe& donlin! o& $%lin!. 8 'al$e a#o'e 20 &e%&esents aea!e& $%lin! and st&onge& donlin!, he&eas 8 'al$e #elo00 o$ld &e%&esent a ea!e& donlin!.
Bf M;A1MA is $sed o& &ecei'e di'e&sit* is a%%lica#le,anadditional 3 d gain is int&od$ced in the $%lin!. Bn s$ch case a
de'iation of 20 is acce%ta#le, i.e, a 8 of 9" o$ld #e no&malin s$ch case.
• 8ath alance Bf the 8 statistic indicates %&o#lem in the
donlin!1$%lin! the - %ath sho$ld #e t&aced fo& %ossi#leha&da&e fa$lts. 8ossi#le things that co$ld go &ong a&e
a) ;igh T:- d$e to fa$lt* feede& ca#le
#) Bm%&o%e& connecto&isation
c) a$lt* com#ine&
Hardware Optimization -Hardware Optimization -Typical Hardware Problems
-
8/8/2019 Sudarshan_ RF Optim
98/186
d) a$lt* antenna im%&o%e& im%edance matching #eteen
antenna and feede& ca#le (&a&e case)
• 8&ocesso& %&o#lems
• he %&esent e$i%ment a&chitect$&e is $ite &o#$st and ith
the e'ol$tion of TLB techni$es, the diffe&ent ha&da&e mod$les
ha'e #een com%acted into single $nits.
• he c$&&ent -Ns1-Hs a&e ha'ing in#$ilt %&ocessing a#ilities a%a&t
f&om also containing the - %h*sical channels.
• ;oe'e& in %laces he&e olde& e$i%ment (fo& e.g. Moto&ola
Bn@ell1Mcell) a&e still in $se, %&o#lems ith %&ocesso& (8-@ o&M@H), co$ld #e enco$nte&ed.
• hese %&o#lems a&e easil* identifia#le #* d&i'e test and $s$all* also
sho $% deg&adation on M@- statistics. ;oe'e& in the c$&&ent
scena&io these %&o#lems ha'e &a&e occ$&ences.
Hardware Optimization -Hardware Optimization -Typical Hardware Problems
-
8/8/2019 Sudarshan_ RF Optim
99/186
• @1&anscode& 8&o#lems Altho$gh the occ$&&ence is &a&e, the&e
a&e instances he&e some %a&t of &anscode& o& timeslot on the
8@M lin! go fa$lt*. Bn s$ch cases, the timeslot ma%%ing needs to #eidentified and a%%&o%&iate t&o$#leshooting ste%s need to #e ta!en.
hese %&o#lems can seldom #e identified #* d&i'e testing.
• te%s fo& ;a&da&e %timization
a) @hec! f&om M@- statistics fo& indications of ha&da&e fa$lts
#) @hec! e'ent logs f&om M@- to find o$t if an* ala&ms e&e
gene&ated
c) @ond$ct call test on the site1cell in $estion chec! fo&
assignment fail$&es, hando'e& fail$&es, f&om la*e& 3 messages.
Hardware Optimization –Hardware Optimization –Hardware Optimization Steps
-
8/8/2019 Sudarshan_ RF Optim
100/186
• te%s fo& ;a&da&e %timization
d) Bsolate the %&o#lem to the s%ecific -N. his can #e done #*
Oloc!ingP the s$s%icio$s -N.e) @hec! fo& donlin! &ecei'e le'el on each -N. Bn some cases the
donlin! &ecei'e le'el on a %a&tic$la& -N ma* #e 'e&* lo, d$e to
fa$lt* &adio.
f) -e$est T:- test to #e %e&fo&med if the %&o#lem a%%ea&s to
#e &elated to %oo& %ath #alance.
g) @hec! fo& im%&o%e& connecto&ization, im%&o%e& antenna
installation. ne loose connecto& co$ld s!e the %e&fo&mance of the
enti&e cellGGG
f) Bf the %&o#lem is not isolated to a #ad -N1 othe& ha&da&e f$&the& in'estigations needed to chec! othe& %ossi#le
fa$lt* ha&da&e in the @1N@C-
-
8/8/2019 Sudarshan_ RF Optim
101/186
Physical RF OptimizationPhysical RF Optimization
-
8/8/2019 Sudarshan_ RF Optim
102/186
• A ell designed - is !e* to good neto&! %e&fo&mance.
• Mo&e often than not, the act$al neto&! #$ilt is de'iated
f&om the neto&! designed f&om the des!to%. he 'a&iationsa&e
a) Act$al site locations a&e aa* f&om the nominal %lannedlocations.
#) Bt is not %&actica#le to #$ild a g&id/#ased neto&! d$e
to se'e&al const&aints.
c) Antenna heights ma* diffe& f&om the %lanned antennaheights.
• 8h*sical - o%timization ma* #e done at se'e&al stages of
neto&! &ollo$t.
Physical RF OptimizationPhysical RF Optimization
-
8/8/2019 Sudarshan_ RF Optim
103/186
• 8h*sical - %timization is an essential &e$i&ement d$&ingthe neto&! #$ild1%&e o%timization stages. Bn most cases the
6M 'endo& is &es%onsi#le fo& the neto&! d$&ing this %haseand he ca&&ies o$t the %&ocess to ens$&e that the act$alneto&! is as nea& good as the des!to% designed one.
• he %&ocess com%&ises of cond$cting a d&i'e test fo& theenti&e cl$ste&, hich ma* com%&ise of one o& se'e&al @
a&eas.• he d&i'e test &es$lts a&e %lotted on a B ma% and
deficiencies in co'e&age1inte&fe&ence %&o#lems a&e identified#* %lotting -7le'1-7$al 'al$es.
• Most of the co'e&age deficiencies a&e fi7ed #* ma!ingchanges to antenna heights(&a&e), #o&e and tilts.
• At late& stages %a&amet&ic o%timization is done to #&ing theneto&! %e&fo&mance close to des!to% design.
Physical RF OptimizationPhysical RF Optimization
-
8/8/2019 Sudarshan_ RF Optim
104/186
• - o%timization is also ca&&ied o$t d$&ing neto&! e7%ansion%hase, i.e hen ne site o& g&o$% of sites a&e added into the
neto&!.• Bn man* neto&!s - o%timization is also done as a &eg$la&
%&ocess to maintain good neto&! %e&fo&mance.
• - o%timization is hel%f$l in &esol'ing s%ecific co'e&age%&o#lems o& inte&fe&ence %&o#lems, cell o'e&&each, no
dominant se&'e& iss$es, etc.• *%ical th$m# &$le to follo hile ca&&*ing o$t %h*sical -
o%timization fo& &esol'ing co'e&age o& inte&fe&ence iss$es /te% D/ &* tilting the antennas.
te% 2D/ &* changing the o&ientation.te% 3D/ Bnc&ease o& &ed$ce the height ifftilt1&eo&ientation does not sol'e the %&o#lem
te% 4D/ @hange the antenna t*%e as a last &eso&t.
-
8/8/2019 Sudarshan_ RF Optim
105/186
Database/Parameter OptimizationDatabase/Parameter Optimization
-
8/8/2019 Sudarshan_ RF Optim
106/186
• he %&ocess sta&ts the moment a M neto&! goes on ai&and contin$es on a da*/to/da* #asis, till the neto&! is
o%e&ational.• Hnde& M each 'endo& has h$nd&eds of %a&amete&s hich
can #e %la*ed ith to achie'e diffe&ent %e&fo&mance met&ics$nde& diffe&ent scena&ios.
• Hs$all* most of the %a&amete&s a&e ena#led ith defa$lt
settings and a&e ala*s !e%t $nchanged. ;oe'e& the&e a&esome s%ecific %a&amete&s hich cont&ol the - %e&fo&mancehich can #e changed on a cell o& e'en ca&&ie&/le'el, toachie'e s%ecific im%&o'ements.
f d d #
Database/Parameter OptimizationDatabase/Parameter Optimization
-
8/8/2019 Sudarshan_ RF Optim
107/186
• M eat$&es efo&e %&oceeding to data#ase %a&amete&s,let $s disc$ss some im%o&tant M feat$&es hich a&e
commonl* #eing $sed in c$&&ent neto&!s.• M neto&!s o&ldide a&e mainl* affected #* the
folloing t*%es of %&o#lemsD/ ) @o'e&age iss$es, 2)Bnte&fe&ence iss$es, 3)@a%acit* iss$es.
• Bnte&fe&ence in M neto&!s can #e &ed$ced significantl*
#* $sing some s%ecial feat$&es, as mentioned • &e$enc* ;o%%ing
• CN and Toice Acti'it* Cetection
• C*namic 8oe& @ont&ol
Database Optimization – Frequency HoppingDatabase Optimization – Frequency Hopping
h i i f h d di d i
-
8/8/2019 Sudarshan_ RF Optim
108/186
• &e$enc* ho%%ing is one of the standa&dised ca%acit*enhancement feat$&es in M s*stem. Bt offe&s a significant
ca%acit* gain itho$t an* costl* inf&ast&$ct$&e &e$i&ements.• &e$enc* ho%%ing can co/e7ist ith most of the othe&
ca%acit* enhancement feat$&es and in man* cases itsignificantl* #oosts the effect of those feat$&es.
• &e$enc* ho%%ing can #e #&iefl* defined as a se$ential
change of ca&&ie& f&e$enc* on the &adio lin! #eteen themo#ile and the #ase station.
• :hen f&e$enc* ho%%ing is $sed, the ca&&ie& f&e$enc* ischanged #eteen each consec$ti'e CMA f&ame. his meansthat fo& each connection the change of the f&e$enc* ma*ha%%en #eteen e'e&* #$&st.
Database Optimization – Frequency HoppingDatabase Optimization – Frequency Hopping
A fi h f h i d i ili
-
8/8/2019 Sudarshan_ RF Optim
109/186
• At fi&st, the f&e$enc* ho%%ing as $sed in milita&*a%%lications in o&de& to im%&o'e the sec&ec* and to ma!e the
s*stem mo&e &o#$st against amming.
• Bn cell$la& neto&!, the f&e$enc* ho%%ing also %&o'ides someadditional #enefits s$ch as f&e$enc* di'e&sit* andinte&fe&ence di'e&sit*.
Database Optimization – Frequency HoppingDatabase Optimization – Frequency Hopping
-
8/8/2019 Sudarshan_ RF Optim
110/186
re/uency
;ime
1
2
3
+all is transmitted through seeralfre/uencies in order toC average the interference (interference diversit')
C minimise the impact of fading (fre&uenc' diversit')
Database Optimization – Frequency HoppingDatabase Optimization – Frequency Hopping
h t th d f f h i i M
-
8/8/2019 Sudarshan_ RF Optim
111/186
• he&e a&e to methods of f&e$enc* ho%%ing in M,=ase#and 2&e$enc* ;o%%ing ( ;) and ,*nthesised2&e$enc* ;o%%ing
(- ;).• Bn the #ase#and f&e$enc* ho%%ing the -Ns o%e&ate at
fi7ed f&e$encies.
• &e$enc* ho%%ing is gene&ated #* sitching consec$ti'e#$&sts in each time slot th&o$gh diffe&ent -Ns acco&ding to
the assigned ho%%ing se$ence.• he n$m#e& of f&e$encies to ho% o'e& is dete&mined #* the
n$m#e& of -Ns
Database Optimization – Frequency HoppingDatabase Optimization – Frequency Hopping
h fi t ti l t f th @@; -N i t ll d t h it
-
8/8/2019 Sudarshan_ RF Optim
112/186
• he fi&st time slot of the @@; -N is not alloed to ho%, itm$st #e e7cl$ded f&om the ho%%ing se$ence.
•
his leads to th&ee diffe&ent ho%%ing g&o$%s.• he fi&st g&o$% doesnPt ho% and it incl$des onl* the @@;
time slot.
• he second g&o$% consists of the fi&st time slots of the non/@@; -Ns.
• he thi&d g&o$% incl$des time slots one th&o$gh se'en f&ome'e&* -N.
Database Optimization – Baseband HoppingDatabase Optimization – Baseband Hopping
-
8/8/2019 Sudarshan_ RF Optim
113/186
B
RTSL 0 1 2 3 4 5 6 7
TRX-1
TRX-2
TRX-3
TRX-4
f1 B = BCCH timeslot. It does not hop.
f2
f3
f4
Time slot 0 of TRX-2,-3,-4 hop over f2,f3,f4.
Time slots 1... of !ll TRXs
hop over "f1,f2,f3,f4#.
%aseband hopping 1%% F>)
Database Optimization – RF HoppingDatabase Optimization – RF Hopping
-
8/8/2019 Sudarshan_ RF Optim
114/186
• Bn the s*nthesised f&e$enc* ho%%ing all the -Ns e7ce%t
the @@; -N change thei& f&e$enc* fo& e'e&* CMAf&ame acco&ding to the ho%%ing se$ence.
• h$s the @@; -N doesnPt ho%.
• he n$m#e& of f&e$encies to ho% o'e& is limited to >3, hichis the ma7im$m n$m#e& of f&e$encies in the Mo#ile
Allocation (MA) list.
Database Optimization – RF HoppingDatabase Optimization – RF Hopping
-
8/8/2019 Sudarshan_ RF Optim
115/186
B(R5/6
Non-BCCH TRXs are hopping over
the MA-list (f1,f2,f3,...,fn atta!he" to the !ell.
(R5/7
B # BCCH ti$eslot. TRX "oes not hop.
f1,
f2,
f3,
fn
f1,
f2,
f3,
fn
. . . .
Synthesised hopping 1RF F>)
Database Optimization – RF HoppingDatabase Optimization – RF Hopping
-
8/8/2019 Sudarshan_ RF Optim
116/186
• he #iggest limitation in #ase#and ho%%ing is that the n$m#e&
of the ho%%ing f&e$encies is the same as the n$m#e& of-Ns.
• Bn s*nthesised ho%%ing the n$m#e& of the ho%%ingf&e$encies can #e an*thing #eteen the n$m#e& of ho%%ing-Ns and >3.
Database Optimization – Frequency HoppingDatabase Optimization – Frequency Hopping
-
8/8/2019 Sudarshan_ RF Optim
117/186
-3
;>-1
-H
1? 2? 3
Dig. 0
;>-1
;>-2
5S+;+S<
5++H
re/uency
;ime
123
-
8/8/2019 Sudarshan_ RF Optim
118/186
• he @ell Allocation (@A) is a list of all the f&e$enciesallocated to a cell. he @A is t&ansmitted &eg$la&l* on the@@;.
• Hs$all* it is also incl$ded in the signaling messages thatcommand the mo#ile to sta&t $sing a f&e$enc* ho%%ing logicalchannel. he cell allocation ma* #e diffe&ent fo& each cell.
• Bn 8M 900 the @A list ma* incl$de all the 24 a'aila#le
f&e$encies UM 04.0?V.• ;oe'e&, the %&actical limit is >4, since the MA/list can onl*
%oint to >4 f&e$encies that a&e incl$ded in the @A list .
Database Optimization – RF Hopping – MobileDatabase Optimization – RF Hopping – MobileAllocationAllocation
• he MA is a list of ho%%ing f&e$encies t&ansmitted to a
-
8/8/2019 Sudarshan_ RF Optim
119/186
• he MA is a list of ho%%ing f&e$encies t&ansmitted to amo#ile e'e&* time it is assigned to a ho%%ing %h*sical channel.
•
he MA/list is a$tomaticall* gene&ated if the #ase#andho%%ing is $sed.
• Bf the neto&! $tilises the - ho%%ing, the MA/lists ha'e to#e gene&ated fo& each cell #* the neto&! %lanne&.
• he MA/list is a#le to %oint to >4 of the f&e$encies defined
in the @A list• ;oe'e&, the @@; f&e$enc* is also incl$ded in the @A list,
so the %&actical ma7im$m n$m#e& of f&e$encies in the MA/list is >3.
•
he f&e$encies in the MA/list a&e &e$i&ed to #e ininc&easing o&de& #eca$se of the t*%e of signaling $sed tot&ansfe& the MA/list.
Database Optimization – RF Hopping – HSNDatabase Optimization – RF Hopping – HSN
• he ;o%%ing ,e$ence J$m#e& (;J) indicates hich
-
8/8/2019 Sudarshan_ RF Optim
120/186
• he ;o%%ing ,e$ence J$m#e& (;J) indicates hichho%%ing se$ence of the >4 a'aila#le is selected.
•
he ho%%ing se$ence dete&mines the o&de& in hich thef&e$encies in the MA/list a&e to #e $sed.
• he ;Js / >3 a&e %se$do &andom se$ences $sed in the&andom ho%%ing hile the ;J 0 is &ese&'ed fo& a se$entialse$ence $sed in the c*clic ho%%ing.
• he ho%%ing se$ence algo&ithm ta!es ;J and J as anin%$t and the o$t%$t of the ho%%ing se$ence gene&ation is aMo#ile Allocation Bnde7 (MAB) hich is a n$m#e& &angingf&om 0 to the n$m#e& of f&e$encies in the MA/lists$#t&acted #* one.
• he ;J is a cell s%ecific %a&amete&.
Database Optimization – RF Hopping – MAIODatabase Optimization – RF Hopping – MAIO
• :hen the&e is mo&e than one -N in the $sing the same
-
8/8/2019 Sudarshan_ RF Optim
121/186
• :hen the&e is mo&e than one -N in the $sing the sameMA/list the Mo#ile Allocation Bnde7 +ffset (MAB) is $sedto ens$&e that each -N $ses ala*s an $ni$e f&e$enc*.
• 6ach ho%%ing -N is allocated a diffe&ent MAB. MAB isadded to MAB hen the f&e$enc* to #e $sed is dete&minedf&om the MA/list.
• MAB and ;J a&e t&ansmitted to a mo#ile togethe& ith
the MA/list.• he MABoffset (Jo!ia) is a cell s%ecific %a&amete& defining
the MAB-N fo& the fi&st ho%%ing -N in a cell. he MABs
fo& the othe& ho%%ing -Ns a&e a$tomaticall* allocatedacco&ding to the MABste%/%a&amete&
Database Optimization – RF Hopping – MAIODatabase Optimization – RF Hopping – MAIO
-
8/8/2019 Sudarshan_ RF Optim
122/186
"S< Hopping algorithm
-3
@ A HS@
;>-1 0
;>-2 1
;>-3 2
or this ;D
-
8/8/2019 Sudarshan_ RF Optim
123/186
he MABste% is a Jo!ia s%ecific %a&amete& $sed in the MAB
allocation to the -Ns.
•
he MAB fo& the fi&st ho%%ing -Ns in each cell is defined #* thecell s%ecific MABoffset %a&amete&
• MABs fo& the othe& ho%%ing -Ns a&e assigned #* adding theMABste% to the MAB of the %&e'io$s ho%%ing -N
• MAB-N(J) = MABoffset + MABste%(n/)
Sector ;> B HS@
-
8/8/2019 Sudarshan_ RF Optim
124/186
Sector ;> B HS@
-
8/8/2019 Sudarshan_ RF Optim
125/186
:hen - ;o%%ing is de%lo*ed the @@; la*e& is %lanned $singthe standa&d 4N3 o& N3 o& an inte&mediate s$ita#le %atte&n.
•
Ma7im$m %&otection is assigned hile %lanning to the @@;la*e& as it is c&itical to call set$% %&oced$&e.
• o& the @; la*e& the&e a&e mainl* th&ee t*%es of idel*$sed &e$se %atte&ns
• N All secto&s in the neto&! $se a single MA list.
• N3 3 MA lists a&e c&eated. ec A of each cell $ses MAL,
ec $ses MAL2 and ec 3 $ses MAL3
• Ad/hoc1Mi7ed ; M$lti%le MA lists a&e $sed. @an ha'e as
man* MA lists as the n$m#e& of secto&s in the neto&!. he
&e$se is #ased on f&actional loading ith a ma7im$m loadingfacto& of 00 F.
Database Optimization – RF Hopping – LoadingDatabase Optimization – RF Hopping – LoadingFactorFactor
• Loading acto& his is the &atio of no of -N to the no of
-
8/8/2019 Sudarshan_ RF Optim
126/186
Loading acto& his is the &atio of no of -N to the no ofho%%ing f&e$encies in the MA list
•
Loading acto& = Jo of ;o%%ing -N1Jo of &e$encies.• o& eg. Loading facto& = "0 F if the&e a&e 2 -N and 4
ho%%ing f&e$encies.
• Loest %&acticall* achie'a#le loading facto& is 33 Ffo&N3, F fo& N and highest is 00 F .
• Hs$all* 00F loading facto& is $sed in case of ad/hoc -ho%%ing, fo& cells ith highe& config$&ation (>/>/>),hoe'e& fo& loe& config$&ation li!e (2/2/2) "0 Floading facto& co$ld #e $sed.
•
Bn case of ad/hoc ho%%ing the loading facto& can #e%lanned to #e s%ecific to the cell config$&ation.
Database Optimization – DTX & Power ControlDatabase Optimization – DTX & Power Control
• 8oe& cont&ol and the CN a&e standa&d M feat$&es,
-
8/8/2019 Sudarshan_ RF Optim
127/186
8oe& cont&ol and the CN a&e standa&d M feat$&es,hich a&e designed to minimise the inte&fe&ence.
•
hese a&e mandato&* feat$&es in the HL, #$t it is $% to theneto&! o%e&ato& to decide hethe& to $se them o& not.
• CN %&e'ents $nnecessa&* t&ansmissions hen the&e is noneed to t&ansfe& info&mation
• 8oe& cont&ol is $sed to o%timise the t&ansmitted signal
st&ength so that the signal st&ength at the &ecei'e& is stillade$ate.
• hese feat$&es can #e indi'id$all* acti'ated fo& $%lin! anddonlin!.
•
%e&ato&s ha'e #een idel* $sing #oth feat$&es in HLdi&ection mainl* in o&de& to ma7imise the #atte&* life inmo#iles.
Database Optimization – DTX & Power ControlDatabase Optimization – DTX & Power Control
• Bn a non/ho%%ing neto&! these feat$&es %&o'ide some $alit*
-
8/8/2019 Sudarshan_ RF Optim
128/186
Bn a non ho%%ing neto&! these feat$&es %&o'ide some $alit*gain fo& some $se&s, #$t this gain cannot #e t&ansfe&&edeffecti'el* to inc&eased ca%acit*, since the ma7im$minte&fe&ence e7%e&ienced #* each $se& is li!el* to &emain thesame.
• he %oe& cont&ol mechanism doesnPt f$nction o%timall*#eca$se the inte&fe&ence so$&ces a&e sta#le ca$sing chain
effects he&e the inc&ease of t&ansmission %oe& of onet&ansmitte& ca$ses o&se $alit* in the inte&fe&ed &ecei'e&,hich in t$&n ca$ses the %oe& inc&ease in anothe&t&ansmitte& and so on.
• his means that, fo& e7am%le, one mo#ile located in a
co'e&age limited a&ea ma* se'e&el* limit the %ossi#ilit* ofse'e&al othe& t&ansmitte&s to &ed$ce thei& %oe&.
Database Optimization – DTX & Power ControlDatabase Optimization – DTX & Power Control
• Bn a non/ho%%ing neto&! these feat$&es %&o'ide some $alit*
-
8/8/2019 Sudarshan_ RF Optim
129/186
n a n n %% ng n t & t f at$& %& m $a t*gain fo& some $se&s, #$t this gain cannot #e t&ansfe&&edeffecti'el* to inc&eased ca%acit*, since the ma7im$minte&fe&ence e7%e&ienced #* each $se& is li!el* to &emain thesame.
• he %oe& cont&ol mechanism doesnPt f$nction o%timall*#eca$se the inte&fe&ence so$&ces a&e sta#le ca$sing chain
effects he&e the inc&ease of t&ansmission %oe& of onet&ansmitte& ca$ses o&se $alit* in the inte&fe&ed &ecei'e&,hich in t$&n ca$ses the %oe& inc&ease in anothe&t&ansmitte& and so on.
• his means that, fo& e7am%le, one mo#ile located in a
co'e&age limited a&ea ma* se'e&el* limit the %ossi#ilit* ofse'e&al othe& t&ansmitte&s to &ed$ce thei& %oe&.
• Bn a &andom ho%%ing neto&! the $alit* gain %&o'ided #*
Database Optimization – DTX & Power ControlDatabase Optimization – DTX & Power Control
-
8/8/2019 Sudarshan_ RF Optim
130/186
%% g * g % *#oth feat$&es can #e efficientl* e7%loited to ca%acit* gain#eca$se the gain is mo&e e$all* dist&i#$ted among the $se&s.
• ince the t*%ical 'oice acti'it* facto& (also called CNfacto&) is less than 0.", CN effecti'el* c$ts the neto&!load in half hen it is $sed.
• he %oe& cont&ol o&!s mo&e efficientl* #eca$se each $se&
has man* inte&fe&ence so$&ces. Bf, one inte&fe&e& inc&easesits %oe&, the effect on the $alit* of the connection is notse&io$sl* affected. Bn fact, it is %&o#a#le that some othe&inte&fe&e&s a&e dec&easing thei& %oe&s at the same time.h$s, the s*stem is mo&e sta#le and chaining effects
mentioned ea&lie& do not occ$& f&e$entl*.
Database Optimization – DTX & Power ControlDatabase Optimization – DTX & Power Control
-
8/8/2019 Sudarshan_ RF Optim
131/186
"(2@C+ on 1. dBD;> on 2.! dB+ on? D;> on !. dB
"(2@C+ on 1. dBD;> on 2.! dB+ on? D;> on !.$ dB
euse 31? ;E 3'm1h euse 31? ;E !0'm1h
+12 impro)ement
he sim$lated gain of 8@ and CN ith ;.
Database Optimization – DTX & Power ControlDatabase Optimization – DTX & Power Control
• CN has some effect on the -N$al dist&i#$tion.
-
8/8/2019 Sudarshan_ RF Optim
132/186
• Jo&mall* the 6- is a'e&aged o'e& the d$&ation of one
A@@; f&ame lasting 0.4? seconds and consisting of 04CMA f&ames.
• ;oe'e&, fo$& of these CMA f&ames a&e $sed fo&meas$&ements, so that onl* 00 #$&sts a&e act$all*t&ansmitted and &ecei'ed.
• :hen CN is in $se and the&e is no s%eech acti'it*, onl* the#$&sts t&ansmitting the silence desc&i%to& f&ame (BC/f&ame) and the A@@; a&e t&ansmitted.
• :hen the&e a&e %e&iods of no s%eech acti'it*, the 6- isestimated o'e& $st the #$&sts ca&&*ing the silencedesc&i%to& f&ame and the A@@;. his incl$des onl* 2#$&sts o'e& hich the 6- is a'e&aged (s$# $alit*).
Database Optimization – DTX & Power ControlDatabase Optimization – DTX & Power Control
• 6- gets a'e&aged m$ch mo&e effecti'el* hen CN is not
-
8/8/2019 Sudarshan_ RF Optim
133/186
g g *$sed *ielding to a $alit* dist&i#$tion he&e the %&o%o&tion ofmode&ate $alit* 'al$es is enhanced.
• he s$# $alit* dist&i#$tion is ide& than the f$ll $alit*dist&i#$tion, meaning that mo&e good and #ad $alit* sam%lesa&e e7%e&ienced.
• he diffe&ences #eteen f$ll and s$# $alit* dist&i#$tions
a&e la&gest in f&e$enc* ho%%ing neto&!s $tilising lof&e$enc* allocation &e$se, since in that !ind of neto&!s theinte&fe&ence sit$ation ma* #e 'e&* diffe&ent f&om #$&st to#$&st.
• A co$%le of se'e&el* inte&fe&ed #$&sts ma* ca$se 'e&* #ad
$alit* fo& the s$# $alit* sam%le hen the* ha%%en to occ$&in the set of 2 #$&sts o'e& hich the s$# $alit* isdete&mined.
Database Optimization – DTX & Power ControlDatabase Optimization – DTX & Power Control
• he f$ll $alit* sam%le of the same time %e&iod has %&o#a#l*
-
8/8/2019 Sudarshan_ RF Optim
134/186
* % % % *onl* mode&ate $alit* dete&io&ation #eca$se of the #ette&a'e&aging of 6- o'e& 00 #$&sts.
• Bn a &eal neto&! $tilising CN the $alit* dist&i#$tion is ami7t$&e of f$ll and s$# $alit* sam%les.
• he %&o%o&tions of f$ll and s$# sam%les de%end on the s%eechacti'it* facto& also !non as the CN facto&.
• he diffe&ences in the 6- a'e&aging %&ocesses ca$sesignificant diffe&ences in the -NHAL dist&i#$tions. hesediffe&ences sho$ld #e ta!en into acco$nt hen the -NHALdist&i#$tions of neto&!s $tilising and not $tilising CN a&ecom%a&ed.
-
8/8/2019 Sudarshan_ RF Optim
135/186
Database Optimization – DTX & Power ControlDatabase Optimization – DTX & Power Control
• 8oe& @ont&ol hat to o%timize
-
8/8/2019 Sudarshan_ RF Optim
136/186
%
• he %a&amete&s to o%timize in case of %oe& cont&ol a&e the
indo settings.
Database Optimization – DTX & Power ControlDatabase Optimization – DTX & Power Control
-
8/8/2019 Sudarshan_ RF Optim
137/186
Conlin! 8oe& @ont&ol *%ical -7le' :indo settings
Downlink Rxlev (dBm)
B S T x P o w e
- / -/
42
Database Optimization – DTX & Power ControlDatabase Optimization – DTX & Power Control
-
8/8/2019 Sudarshan_ RF Optim
138/186
Downlink Rx!"#l
B S T x P o w e
0 4
42
Conlin! 8oe& @ont&ol *%ical -7$al :indo settings
Database Optimization – DTX & Power ControlDatabase Optimization – DTX & Power Control
-
8/8/2019 Sudarshan_ RF Optim
139/186
$%link Rxlev (dBm)
& S T x P o w e
- 0 -0
33
H%lin! 8oe& @ont&ol *%ical -7le' :indo settings
/
Database Optimization – DTX & Power ControlDatabase Optimization – DTX & Power Control
• 8oe& @ont&ol %a&amete&s hich can #e set
-
8/8/2019 Sudarshan_ RF Optim
140/186
• Conlin!1H%lin! -7le' th&eshold (lS&7le'SdlS% 1lS&7le'S$lS%)
• -7$al th&eshold(lS&7$alSdlS% 1 lS&7$alS$lS%)
• 8oe& inc&ement1&ed$ction ste% size(%oSincSste%SsizeSdl1%oS&edSste%SsizeSd)
• C*namic ste% ad$st algo&ithm(d*nSste%Sad)
• 8oe& @ont&ol eat$&es
Database Optimization – DTX & Power ControlDatabase Optimization – DTX & Power Control
-
8/8/2019 Sudarshan_ RF Optim
141/186
• #ecti'e is to &ed$ce a'e&age inte&fe&ence
•
Bn case of $%lin! also hel%s in sa'ing #atte&* %oe&• Algo&ithm o&!s on meas$&ement &e%o&ts sent #* the M
e'e&* 4?0 ms (A@@; f&ame)
• Conlin! %oe& cont&ol cannot #e a%%lied to @@; ca&&ie&
• H%lin! %oe& cont&ol is mandato&* #$t donlin! %oe& cont&olis not mandato&*. eat$&e selecta#le #* the o%e&ato&.
• o& cont&olling inte&fe&ence in the neto&! the o%e&ato& $sesCN, 8oe& @ont&ol and &e$enc* ;o%%ing. hese feat$&eseffecti'el* act as com#ined fo&ces in inte&fe&ence &ed$ction
and im%&o'ed call $alit*.
Database OptimizationDatabase Optimization
*%ical %&o#lems hich M s$#sc&i#e&s e7%e&ience a&e
-
8/8/2019 Sudarshan_ RF Optim
142/186
• @o'e&age iss$es
•
Toice $alit* iss$es• Access iss$es1congestion
• ;ando'e& &elated iss$es
• C&o%%ed calls
8a&amete&s a&e #&oadl* classified into the folloing g&o$%s
Database OptimizationDatabase Optimization
-
8/8/2019 Sudarshan_ RF Optim
143/186
• Access &elated %a&amete&s
•
@all handling1;ando'e& &elated %a&amete&s• @ongestion &elated %a&amete&s
Database OptimizationDatabase Optimization
-
8/8/2019 Sudarshan_ RF Optim
144/186
Database Optimization – IDLE Mode CellDatabase Optimization – IDLE Mode CellSelectionSelection
W he M $ses a %ath loss c&ite&ion %a&amete& @ tod t min h th ll is s it #l t m% n UM 03 22V
-
8/8/2019 Sudarshan_ RF Optim
145/186
dete&mine hethe& a cell is s$ita#le to cam% on UM 03.22V
W @ de%ends on 4 %a&amete&sDW . -ecei'ed signal le'el (s$ita#l* a'e&aged)
W 2. he %a&amete& &7Le'AccessMin, hich is #&oadcast on the@@;, and is &elated to the minim$m signal that the o%e&ato&ants the neto&! to &ecei'e hen #eing initiall* accessed #* an
MW 3. he %a&amete& ms78&Ma7@@;, hich is also #&oadcast on
the @@;, and is the ma7im$m %oe& that an M ma* $se heninitiall* accessing the neto&!
W 4. he ma7im$m %oe& of the M.
Database Optimization – IDLE Mode CellDatabase Optimization – IDLE Mode CellSelectionSelection
Cell "electionin.D,# *ode8&asedonC6
-
8/8/2019 Sudarshan_ RF Optim
146/186
Cell "election in .D,# *ode8 &ased on C6
9 Radio Criteria
A : Recei)ed ,e)el A)erage / p6
C6 : ;A / *a1;B8
-
8/8/2019 Sudarshan_ RF Optim
147/186
W Bn case of &eselection f&om one cell to anothe& in the same locationa&ea the @ 'al$e of ta&get cell m$st #e highe& than so$&ce cell
W Bn case of &eselection to a ta&get cell in a diffe&ent location a&ea the @ 'al$e m$st #e g&eate& than that of the so$&ce cell #* adata#ase %a&amete& QcellS&eselectSh*ste&esisR
@ell -eselection @2
W @2 is an o%tion M feat$&e hich can onl* #e $sed fo& cell&eselection, it can #e ena#led o& disa#led on a cell #asis.
W Bf @2 %a&amete&s a&e not #eing #&oadcast the @ %&ocess is $sed fo&&eselection.
@ell -eselection @2
Database Optimization – IDLE Mode CellDatabase Optimization – IDLE Mode CellSelectionSelection
-
8/8/2019 Sudarshan_ RF Optim
148/186
W @2= @ + cellS&eselectSoffset tem%o&a&* offset ;
(%enalt*Stime ) (fo& %enalt*Stime E3)W ;= 0 if %enalt*Stime
W ;= if E %enalt*Stime
W @2= @ cellS&eselectSoffset (fo& %enalt*Stime= 3)
:h* @2
-
8/8/2019 Sudarshan_ RF Optim
149/186
W Bn d$al#and neto&!// to gi'e diffe&ent %&io&ities fo&
diffe&ent #andW Bn m$ltila*e&// to gi'e %&io&it* to mic&ocell fo& slo mo'ing
t&affic
W An* othe& s%ecial case he&e s%ecific cell &e$i&ed highe&%&io&it* than the &est
@ell -eselection t&ateg*
W 8ositi'e offset// enco$&age Ms to select that cell
W Jegati'e offset// disco$&age Ms to select that cell fo& the
d$&ation %enalt* ime %e&iod
Database OptimizationDatabase Optimization
-
8/8/2019 Sudarshan_ RF Optim
150/186
Database Optimization – HandoversDatabase Optimization – Handovers
;ando'e&
-
8/8/2019 Sudarshan_ RF Optim
151/186
W he hando'e& (;) %&ocess is one of the f$ndamental%&inci%les in cell$la& mo#ile &adio, maintaining the call in%&og&ess hilst the mo#ile s$#sc&i#e& is mo'ing th&o$gh theneto&!.
W Bn idle mode the M does a cell &eselection, he&eas indedicated mode the M %e&fo&ms a hando'e&.
W ;ando'e&s a&e mainl* classified into to t*%es
W A) Bnte& cell hando'e&s
W ) Bnt&a cell hando'e&s
W Bnte& cell hando'e&s f$&the& classified asW Bnte& ie #eteen to cells #elonging to diffe&ent
@s
W Bnt&a ie #eteen to cells #elonging to same @
;ando'e&
Database Optimization – HandoversDatabase Optimization – Handovers
-
8/8/2019 Sudarshan_ RF Optim
152/186
W Bnt&a cell hando'e&s is the sitching of call f&om onechannel1-N to anothe& -N ithin the same cell1. hisis an o%tional feat$&e hich can #e ena#led on a cell #asis.Bnt&a cell hando'e&s $s$all* ta!e %lace hen the -7$al onthe so$&ce channel dete&io&ates.
;ando'e& %&ocess ma* #e initiated d$e to the folloing main
&easonsW -adio @&ite&ia
W o maintain &ecei'e le'el1&ecei'e $alit*
W A#sol$te M/ distance
W 8oe& $dget
W Jeto&! @&ite&ia
W &affic load (to manage t&affic dist&i#$tion)
W ;ando'e&s also classified as im%e&ati'e1non/im%e&ati'e #asedon the &eason fo& hich the %&ocess is t&igge&ed
Database Optimization – HandoversDatabase Optimization – Handovers
-
8/8/2019 Sudarshan_ RF Optim
153/186
on the &eason fo& hich the %&ocess is t&igge&ed.
W he ca$se 'al$e contained in the hando'e& &ecognisedmessage ill affect the e'al$ation %&ocess in the @.
;ando'e& ca$ses ma* #e %&io&itized as follos
W . H%lin! $alit*
W 2. H%lin! Bnte&fe&enceW 3. Conlin! $alit*
W 4. Conlin! Bnte&fe&ence
W ". H%lin! Le'el
W >. Conlin! Le'el
W . Cistance
W ?. 8oe& $dget
Database Optimization – HandoversDatabase Optimization – Handovers
8oe& #$dget hando'e&
-
8/8/2019 Sudarshan_ RF Optim
154/186
W Bf an M on a allocated &eso$&ce d$&ing its meas$&ement&e%o&ting %&ocess sees anothe& channel that o$ld %&o'ide ane$al o& #ette& $alit* &adio lin! &e$i&ing a loe& o$t%$t%oe& then a hando'e& ma* #e initiated.
W ;ando'e&s d$e to %oe& #$dget ens$&e that the M is ala*slin!ed to the cell ith minim$m %athloss tho$gh the $alit*
and le'el th&esholds ma* not #e e7ceeded.W ;ando'e& to the ta&get cell ta!es %lace hen 8
hoMa&gin8
W 8 = (ms78&Ma7 A'S-7le'SCLS; (#ts78&Ma7
SN8:-)) (ms78&Ma7(n) A'S-7le'SJ@6LL(n))he&e n RnthR adacent cell hich is a hando'e& candidate
Database Optimization – HandoversDatabase Optimization – Handovers
8oe& #$dget hando'e&
-
8/8/2019 Sudarshan_ RF Optim
155/186
W hoMa&gin8 is a %a&amete& hich can #e set on a cell tocell #asis. 6ach cell ma* ha'e a diffe&ent 'al$e fo& eachneigh#o$& cell hich is a candidate fo& %oe& #$dgethando'e&.
W hoMa&gin is e7%&essed in d and is $s$all* set to 4. ;oe'e&this ma* #e &ed$ced if the hando'e& needs to #e s%eeded o&
inc&eased to > o& highe& to %&e'ent %ing/%ong o& to dela*hando'e&s
W Bn some cases negati'e homa&gin ma* also #e $sed.
Database Optimization – HandoversDatabase Optimization – Handovers
;ando'e& Algo&ithms
d l h d dd d f l
-
8/8/2019 Sudarshan_ RF Optim
156/186
W ;ando'e& algo&ithms a&e $sed in addition to defa$lt%a&amete&s to cont&ol the hando'e& %&ocess
W hese algo&ithms assist in mo#ilit* management and a&eeffecti'e in t&affic dist&i#$tion.
W he algo&ithms ha'e an im%o&tant &ole to %la* in Mneto&!s hich $se m$lti/#and o& m$lti/la*e& a&chitect$&es.
;ando'e& Algo&ithms
M l h d h #
Database Optimization – HandoversDatabase Optimization – Handovers
-
8/8/2019 Sudarshan_ RF Optim
157/186
W Bn Moto&ola s*stem the&e a&e %&oced$&es. hese a&e set #*the %a&amete& %#gtSalgSt*%e. he algo&ithms a&e #&iefl*defined as follosD/
W *%e @on'entional M 8
W *%e 2 -est&icted 8 fo& mac&o cells
W *%e 3 8 ith -7le' as $alifie&W *%e 4 8 ith time in cell as $alifie&
W *%e " 8 ith dela* since neigh#o$& le'el e7ceedsth&eshold as $alifie&
W *%e > Cela*ed %oe& #$dget $sing d*namic hando'e&ma&gin
W *%e 8 algo&ithm to a'oid adacent channelinte&fe&ence
;ando'e& Algo&ithms
f h h l d 2 3 d
Database Optimization – HandoversDatabase Optimization – Handovers
-
8/8/2019 Sudarshan_ RF Optim
158/186
W f the se'en, the most commonl* $sed a&e *%e, 2, 3 and .
W 6ach hando'e& candidate cell can #e defined as one of these'en t*%es of neigh#o$& to the so$&ce cell.
;ando'e& %e& ca$se
W he hando'e& %e& ca$se statistic is a co$nte&/a&&a* statistic
hich co$nts the &eason fo& each hando'e& e'ent on all cellsfo& hich it is ena#led.
W his statistic gi'es im%o&tant info&mation a#o$t the hando'e&%e&fo&mance of the cells and can #e $sed fo& t&o$#leshootingcells hich ha'e high Qhando'e& fail$&e &ateR.
;ando'e& %e& neigh#o$&
hi t ti ti i th l f f h d tt t
Database Optimization – HandoversDatabase Optimization – Handovers
-
8/8/2019 Sudarshan_ RF Optim
159/186
W his statistic gi'es the 'al$e of no of hando'e& attem%ts asell as s$ccesses fo& each neigh#o$& cell. his statistic isalso hel%f$l in t&o$#leshooting hando'e& %e&fo&mance, it can#e $sed to identif* neigh#o$& &elations hich ha'e a highQhando'e& fail$&e &ateR
W he hando'e& %e& neigh#o$& statistic can also #e $sed fo&
neigh#o$&list %&$ning.
Database OptimizationDatabase Optimization
-
8/8/2019 Sudarshan_ RF Optim
160/186
Database Optimization – TRHO/CongestionDatabase Optimization – TRHO/CongestionRelated ParametersRelated Parameters
-; :hat does it do
-
8/8/2019 Sudarshan_ RF Optim
161/186
W -; effecti'el* &ed$ces the se&'ice a&ea of the congestedcells
W Bnc&eases se&'ice a&ea of $nde&/$tilised ta&get cells
W ; is t&igge&ed $sing a s%ecial %a&amete&Qamh&ho8#gtMa&ginR instead of hoMa&gin8#gt
W ene&al g$idelineDW a&get cell Q-7le'accessminR sho$ld #e set highe& to
a'oid #ad donlin! -7$al afte& ;
W amh&ho8#gtMa&gin m$st #e loe& than hoMa&gin8#gt
Database Optimization – TRHO/CongestionDatabase Optimization – TRHO/CongestionRelated ParametersRelated Parameters
-;1@ 8a&amete&s
-
8/8/2019 Sudarshan_ RF Optim
162/186
W amhH%%e&loadth&eshold his %a&amete& dete&mines
minim$m t&affic load th&eshold at hich cell sta&ts to intiate-; defa$lt 'al$e ?0 F
W amhMa7Loadfa&get@ell his %a&amete& dete&minesma7im$m t&affic load th&eshold #e*ond hich ta&get cell illnot acce%t -; hand/ins defa$lt 'al$e >0 F
-;1 8a&amete&s
W amh&ho8#gtMa&gin his %a&amete& is ne 8#gt ma&ginhen cell e7ceeds amhH%%e&loadth&esh. BtPs the &e'ised%oe& #$dget ma&gin hich &e%laces the no&mal 8#gt
definition hen the &ho c&ite&ia a&e met defa$lt 'al$e is "d.
Database Optimization – TRHO/CongestionDatabase Optimization – TRHO/CongestionRelated ParametersRelated Parameters
-;1Adacenc* 8a&amete&s
-
8/8/2019 Sudarshan_ RF Optim
163/186
W t&hoa&getLe'el his %a&amete& dete&mines the minim$m
-B of the 'alid ta&get cell candidate &e%o&ted #* themo#ile defa$lt is ?" dm
Ci&ected -et&*
Database Optimization – TRHO/CongestionDatabase Optimization – TRHO/CongestionRelated ParametersRelated Parameters
-
8/8/2019 Sudarshan_ RF Optim
164/186
W A t&ansition (hando'e&) f&om C@@; in one cell to a @; in
anothe& cell d$&ing call set$% d$e to $na'aila#ilit* of an em%t*@; ithin the fi&st cell.
W o cont&ol t&affic dist&i#$tion #eteen cells to a'oid a call&eection.
W @an #e $sed fo& #oth M@ and M@W etting g$idelinesD
W d&h&eshold sho$ld #e highe& than -7le'mincell(-7le'accessmin)X else the im%&o'ed ta&get cell selectionc&ite&ia ill #e igno&ed.
@ongestion -elief
Database Optimization – TRHO/CongestionDatabase Optimization – TRHO/CongestionRelated ParametersRelated Parameters
-
8/8/2019 Sudarshan_ RF Optim
165/186
W his %&oced$&e is initiated hen an M is assigned to an
C@@;, &e$i&es a @; and none a&e a'aila#le.W o o%tions a&e offe&ed fo& deciding ho man* hando'e&
%&oced$&es a&e act