Top Banner

of 5

WCDMA Downlink Coverage, Interference Margin for Users Located at the Cell Coverage Border

Jun 04, 2018

Download

Documents

Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
  • 8/13/2019 WCDMA Downlink Coverage, Interference Margin for Users Located at the Cell Coverage Border

    1/5

    WCDMA Downlink Coverage: Interference Margin for Users Located at the CellCoverage Border

    Kimmo Hiltunenand Nicola Binucci*Oy LM E ricsson Ab, Telecom R&D, FIN-02 420 JORVAS, Finland*Encsson Telecomunicazioni SPA, I40033 Casalecchio di Reno BO), Italytel: +358 9 2993309, email: [email protected]

    Abstract The scope of this paper is to derive a simple equationfor the WCDMA downlink interference margin as a function ofthe relative downlink load. This interference margin can then beused as an input for a typical link budget. Due to the complexityof the WCDMA downlink, a fully analytical approach wouldprobably result in a rather complicated andlor inaccurateapproximation. Therefore, in this paper the actual equation forthe interference margin is derived from the WCDMA theory,and the missing parameter values are obtained from a number ofsystem simulations. The downside of this approach is that a fullyflexible usage of the approximated interference margin requiresinput from a number of measurements or system simulationswith different initial assumptions.

    I. INTRODUCTIONIn a WCDMA system, the link budget can typically be

    Pnlm is the maximum transmit power,p, is the required carrier-to-interference atio (CIR) atthe cell coverage border, including the loss due tothe lac k of fast power control, and the pos siblemacro diversity gain,is the thermal noise power o f the receiver,is the in terference margin app licable for the cellcoverag e border,is the chip rate (3.84 Mcps),

    NIWR is the bit rate.As can be noticed, one of the key inputs for the W CDMAlink budge t is the interference mTg in, wh ich tells how muchthe total interferen ce level experienced by a user at the cellcoverage border is increased due to the impact of multipleusers c omm unicating at the same time.

    constructedas In the uplink, a well known relationsh ip exists between heuplink relative load level q u ~nd the uplink interferenceA + B . lOlog(d) 1) marginIuL:

    (4) tz+ G R B S + G J E - M s , - P + N + r

    I , =which, when re-organized, and assuming that the cell I - r l U LThe scope of this paper is to highlight some of the keycoverage is estimated, can be written as

    (3)In the eq uations given above:A is the attenuation constant,B is the attenuation factor,d, is the cell range,L d dGms is the base station antenna gain,GUE is the m obile station antenna gain,M,I, is the shad ow fading margin for the requiredcoverag e probability level, including the po ssible

    soft hando ver gain,

    are the a dditional losses, e.g. feeder loss and bodyloss,

    0-7803-7484-3/02/$17.00 02002 IEEE. 270

    characteristics of the WCDMA downlink coverage, and in theend, to derive a simple equation for the WCD MA downlinkinterference margin IDL as a function of the relative downlinkload qDb Due to the complexity of the W CDM A dow nlink, afully analytical approach would probably result in a rathercomplicated andor inaccurate approxim ation. Therefo re, inthis paper, a semi-ana lytical approach is chosen: First, inChapter I1 the actual equation for the downlink interferencemargin as a function of the relative downlink load is derivedfrom the WC DMA theory. Then, in Chap ter I a fewexamples are given using parameter values obtained from anumber of simple system simulationsasan nput.II .THEORY I

    A . Downlink Integerence MarginWhen the WCDMA downlink is considered, the totalinterference consists of three parts: intracell interference,intercell interference and the thermal noise. Traditionally, the

    VTC 2002

  • 8/13/2019 WCDMA Downlink Coverage, Interference Margin for Users Located at the Cell Coverage Border

    2/5

    intracell interference is expressed as a certain fraction a f thetotal received power from the own base station. Furthermore,one can also assume that the total intercell interference is Ftimes the total received power from the own cell. Thus, ingeneral, the downlink interference margin (noise rise) formobile i ZDL an be calculated as

    and further, when the path loss L;,, is solved from thedownlink CIR equation:

    In 6) pi is the required CIR including the impact ofmultipath fading and the possible macrodiversity gain,is the downlink orthogon ality factor,is the inter-to-intracell interfere nce ratio,is the downlink transmit power for user i,

    a,F jP,,,Plot,, is the total base station o utput power.

    Assuming now that user i models the typical behavior of auser located at the cell coverage border, e.g. in case of 95area coverage the behavior of the 95th percentile of the usersthe downlink interference margin becomes

    maxwhere a, s the downlink orthogonality factor and F, s theinter-to-intracell interference ratio for the location of the cellcoverage border.B. Total Base Station Output Power

    In general, the total base station output power consists ofthe power allocated for the dedicated channels (Pi,,),akingthe soft handover overhead into account, and the powerallocated for the common channels P C C H , ~ ) .husK

    i Iwhere vi is an activity factor. Furthermo re, the averag enumber of radio links connected to a base station can becalculated as

    ASK = M x eZ.

    x = l9)

    In 9) it is assumed that the average number of users percell is M, he maximum size of the active set is A S and E isthe probability that the active set has n inks.The required transmit power towards a certain user can besolved from the dow nlink CIR equa tion

    Plf l = pl ai,mFl,)Ptot,,,, + NL,,J 10)= p,((a,, + Flm)p* ot,m NkimLc)

    where k; is the ratio between Li, , and LeThe required total common channel power, Pcc~,,depends obviously on the cell size, but also on the cell load.The dependency on the cell load is twofold: Firstly, therequired peak powers depend on the Ptot,m n a similar way as

    the dedicated channels, see 10). Secondly, the activityfactors for some of the dow nlink common channels depend onthe number of users due to the fact that the channels inquestion are used to can y information towards specific users.Such physical channels include for example the SecondaryCommon Control Physical CHannel (S-CCPCH), PagingIndicator CHannel (PICH) and Acquisition Indicator CHannel(AICH). On the other hand, the average po wer transmitted onthe Common Pilot CHannel (CPICH), the Primary CommonControl Physical CHannel (P-CCPCH) and theSynchronization CHannel (SCH) d oes no t depend on the cellload [l].In this paper, the average total power on the commonchannels is modeled so that the required peak powers dependon the cell size, whereas the activity factors depend on the cell

    load, if applicable. Furthermore, it is assumed that the ratiosbetween the required peak po wers are constant. Therefore, itis enough to define the required peak transmit power for onecommon channel, CPICH in this paper, and the transmitpowers for all the others can then be easily derived fromIn the simplest form one can assume that the ratio betweenP c c H and PCPICH s linear as a function of the averagenumber of users per cell. Thus,

    PCPICH.

    P C C H ~ Cc C I M ) P C P I C H ~= YCC, + C,MN a, + Fc)Pt0rp+NL,)

    11)

    where the PCPIcH,, is solved from the CPICH CIR equationfor the cell coverage border. Furthe rmore , in (1 1):C,C

    is the average total load-invariant common channelpower (relative to the C PICH p ower),is the average total (normalized) load-variant

    0-7803-7484-3/02/ 17.00 02002 IEEE. 27 VTC 2002

  • 8/13/2019 WCDMA Downlink Coverage, Interference Margin for Users Located at the Cell Coverage Border

    3/5

    common channel power (relative to the CPIC Hpower and a reference load level,Mo ,is the required CIR for the CPICH.

    Now, assuming a single service system, and following theapproach presented in [2], the users c n be divided intodifferent subgroups based on the number of active links theyhave. Hence, the average total base station output powerbecomesA S

    PtoI = P,,+M. X E ~ V ~ P ~ [ P ~ ~ ~ ~ ~ + ~ ~ ) + N ~ ~ ~ ~ I12)x =

    where px is the average effective CIR target,V~ rF,k,

    is the average activity factor,is the average downlink ortho gonality factor,is the average inter-to-intracell interferenceratio, andis the average ki,nl actor valid for grou p x.Solving (12) for Pror nd letting

    PtOI 0 (13)results in the following downlink pole capacity Mpole

    The Pro,which can be solved from (12)is he averag e totalbase station output power fora certain cell size. However, forthe link budget calculations he path lossL, should depend onthe maximum allow able Pma valid for the service in question.This relationshipcanbe obtained by assu ming that

    Finally, taking both (12) and 15) into account, the Prorcanbe solved as

    where

    andIliJL = pole (17)

    C. Summaty a nd LimitationsThe final expression for the WCDMA downlinkinterference margin can be obtained by comb ining 7) and16). Thus, the IDLbecomes

    As can be noticed , the IDLdoes not depend on the Pmnr ndpc. However, as a result of the fact that in reality the Pror slimited to some maximum value, PBS,,,, the relative cell loadDL will have a certain maximum value, which depends onthe chosen PBsmnr,Pmerand pCIn general, the WCDMA downlink coverage can be limitedeither by the P- or by the PBS,,. However, since theoutcom e of a traditional link budget is the maximum cov eragewith a certain P,, one of the basic assumptions in the

    approach presented in this paper is that the base stationoverload probability, i.e. th e probability that the req uired totalbase station output power exceeds PBsm, is small comparedto the required outage probability. This assum ption can betaken into account by leaving a special margin, headroomR , etween the PBsmar nd the m aximum allowable P o t . Thesize of the required headroom will depend among other thingson the service and the relative cell load [3].Assuming that PIoI BSmax R the upper boundary for theDL can be solved from (1 6)

    aBSmaxI - - - PCCY 20)RL 2p s + l -=J - ac+F,) s+ C,

    where the normalized maximum link power is defined as

    However, it should be kept in mind that at all times0 nor e 1 . This can be also verified by lo oking at the L, as afunction of the q D L , see (22). There, one can easily noticethat when qDr+ 1 , L , + 0. In practice, the plannedmaximum T DL should not be too large, e.g. above 0.8, even if(20) would allow a larger value. This is due to the fact that athigh average load levels a modest change in the number ofusers can cause a considerable increase in the required basestation ou tput power.

    An altem ative way to ap ply the interference margin is to

    0-7803-7484-3/02/ 17.00 02002 IEEE. 272 VTC 2002

  • 8/13/2019 WCDMA Downlink Coverage, Interference Margin for Users Located at the Cell Coverage Border

    4/5

    assume a certain relative load level qDL,n which case theupper boundary for the normalized maximum link powerbecomes

    a,F

    Instead of using a special headroom, the impact of trafficvariations can be taken into account already in the Ptol value[2]. For example, a small margin can be added into p and P sothat when PtOt= PBsmar,he b ase station ov erload probabilitydoes not exceed the requ ired level.There exist also a number of limitations for theapplicability of the derived interference margin. For example:

    Parameters Mpoio S, pc and Fc depend typically notonly on the service, but also on the chosen environ-ment, channel model, user speed, cell plan and softhandover parame ter setting.Different traffic distributions have also an impact onthe MpOloS pc and Fc parameters.The derived equation is not accurate for large outageprobabilities. The reason behind this is that the Ptotexpression is valid only for unlimited link and totalbase station output powers.

    Thus, as a summary of the above; a fully flexible usage ofthe derived approximation for the downlink interferencemargin requires results from a number of measurements orreference simulations with different initial assumptions.Furthermore, the approximation is suitable only for relativelylow outa ge probabilities.

    0.642.1

    111. EXAMPLEIn this chapter, a simple example is given for two different

    types of services: 12.2 kbps speech, and 128 kbps circuitswitched data. A single service system with fully coveredservices is assumed. The users are assumed to be uniformlydistributed (in averag e) over the who le system area. Theenvironment is assumed to be urban, and the cell plan isassumed to consist of hexagonal 3-sector cells with 65degreeantenna lobes.Assume first that the reference simulations for the wantedsystem scenario have resulted in the input values listed inTable I. Note that v is the average activity factor and p is thetraditional downlink CIR target given in desibels.Furthermore, the parameters listed in Tab le I1are assumed to

    be valid for the downlink common channels. Finally, theservices are assumed to have the characteristics given in Table111.

    TABLE Ih I P T FROM THE REFERENCE SIMULATIONS.r IParameter Value I

    1.61 .v .I

    P 0.29. v .I I

    TABLE I1INPUT P A R A M ~ E R SOR THE DOWNLINKOMMONHANNELS

    Parameter Value

    1.4125 = 0.056-16 dB

    TABLE 111SERVICE CHARACTERISTICS.

    I 12.2kbps 7 d B -18.0dB 0.67 II 128kbps 5dB -9.8dB 1.0 IThus, now the values for the Mpoleand S can be calculated,see Table IV. Furthermo re, the correspond ing unconstraineddownlink interference margins as a function of the relativedownlink load are drawn in Fig. 1.

    TABLE IVINPUT PARAMETERS OR THE INrrWERENCE bfAFX3E-4.

    Service Mpore0.188

    128 kbps 0.161

    0-7803-7484-3/02/ 17.00 02002 IEEE. 273 VTC 2002

  • 8/13/2019 WCDMA Downlink Coverage, Interference Margin for Users Located at the Cell Coverage Border

    5/5

    Fig. 1. Downlink interference margin a s a function of therelative downlink load.Assuming for simplicity reasons that R = 0.2. PBsmax forthe 12.2 kbps speech and i2 = 0.3 - P B S m a x or the 128 kbpsdata service the qmax p)an be drawn as shown in Fig. 2.

    Fig. 2. The up per boundary fo r the relative downlink loadqDL)s a function of the norm alized maximum link powerP).

    Finally, an example on the maximum path loss L, as afunction of the relative downlink load qDL s given for anumber of differentp values, see Fig. 3. There, the 12.2 kbpsspeech service is assumed . Furtherm ore, the thermal noisepower N s assumed to b e equal to -100 dBm and the PBS,,, isassumed to be 2 0 W. Note also, that the limitation given in(20) is taken into acco unt.

    RUhaoadmM

    Fig. 3. Maxim um path lossL, as a function of the relativedownlink load for differentp values.IV. CONCLUSIONS

    In this paper a se mi-analytical expression for the downlinkinterference margin as a fu nction of the relative downlink loadis presented. The actual equation is derived from theWCDMA theory, while the missing parameter values areassumed to be obtained with m easurementsor simple systemsimulations.The results show that the expression for the downlinkinterference margin is not as simple as the correspondingexpression for the uplink. On top of that, the requiredparameter values depend on numerous factors, e.g. service,environment, channel model, user speed and soft handoverparameters. Therefore, a fully flexible usage of the derived

    interference margin requires results from a number ofmeasurements or reference simu lations with different initialassumptions.REFERENCES

    [l] 3GPP TSG RAN 25.211 V3.9.0, Physical channels andmapping of transport channels onto physical channels(FDD) (Release 1999) , December 2001.[2] K. Sipilti, 2-C . Honka salo, J. Laiho-Steffens, A. Wacker,Estimation o f capacity and required powe r of WCD MAdown link based on downlink pole equation , P roceed-ings of VTC 2000 (spring ), pp. 1002-1005.[3] K. Hiltunen, R. DeBemardi, 'WCDMA downlink capac-

    ity estimation , P roceedin gs of VTC 2000 (spring), pp.992-996.

    0-7803-7484-3/02/ 17.00 02002 IEEE. 274 VTC 2002