Power control for interference management in LTE femto ... PC for interference mgmt in LTE Femto Zhao Zhao, Thomas Kaiser 18/02/2010 2. Why Femto at home? How many mobile phone calls

Power control for interference management gin LTE femto cell

An overview

Zhao Zhao, Thomas KaiserIKT Uni HannoverIKT, Uni Hannover

[email protected]

Presentation OutlineIntroduction to femto cell

I t f bl i LTE f t llInterference problems in LTE femto cell

Power control for interference managementPower control for interference managementdownlink power controlsuplink power controlsadvanced power controls

Preliminar sim lation res ltsPreliminary simulation results

Concluding remarksConcluding remarks

PC for interference mgmt in LTE FemtoZhao Zhao, Thomas Kaiser

18/02/2010

2

Why Femto at home?How many mobile phone calls are made at Home?

Mobile Phone Usage in USA 20071 “Femto Cell Market to Reach $630

39%

32%

25%30%35%40%45% Million in 2010” 3

14%

4%

11%

5%10%15%20%25%

58%60%

70%

Mobile Phone Usage at Homein USA2

0%At Home In Transport At Work Traveling Others

39%43%

30%

40%

50%

60%

Femto cell is one promising solutions for mobile coverage at home:

0%

10%

20%

2007 2008 2013(est.)

mobile coverage at home:Reliable communicationEasy installLow price and OPEX

1 – “The Case for Home Base Stations” tech. white paper by Femtoforum, Apr. 11, 20072 – “Mobile Phone Use in the Home is Growing” by David H. Deans, Sep. 29, 20083 – “Mobile market research report” by Infonetics Research, 2007

Seamless handoverNetwork integration


18/02/2010

3

Femtocell for LTE

eNB- evolved Node B; UE- User Equipment; HeNB- Home evolved Node B;


18/02/2010

4

Femto cell and spectrum efficiencyWe are running out of spectrum!

Area spectrum efficiency

1214161820

Area spectrum efficiency

02468

1012

GSM EDGE Wimax

WiFi (802.1

1n)

HSDPA LTE Femto

*

bps/Hz per site 0.17 0.33 1.2 2.4 2.88 16.32 18

0

*

Since 1957, the wireless capacity has an approximately million fold increase, 25ximprovement from wider spectrum, 5x improvement from dividing spectrum into smallerslices, 5x improvement by designing better modulation schemes, and a whopping 1600x

* - current industry goals

, p y g g , pp ggain through reduced cell sizes and transmit distance.”2

2 – “Area Spectral Efficiency of Cellular Mobile Radio Systems”, by M. –S Alouini and Andrea Goldsmit, IEEE Trans. Vehic. Tech. Jul. 1999, pp. 1047-66


18/02/2010

5

What challenges Femtocell will bring?Technical challenges

Interference problems Low cost implementationLow cost implementationNetwork and integrationSeamless handover

Non-technical challengesL ti ditLow operating expenditureSecurity problemsLawful issuesHealth care problems


18/02/2010

6

Interference problems in LTE FemtocellFemtocell number per macro cell is indefiniteFemtocell location can be without planningF t t f LTE i l d iFrequency reuse strategy for LTE is already aggressiveLTE air interfaces: OFDMA in downlink and FC-FDMA in uplinkRestrictions on the functionalities of Home eNBRestrictions on the functionalities of Home eNBImperfect synchronization for FemtocellsPriority issuesPriority issuesClose/Open Connectivity issues

Efforts from both academy and industries: esp. 3gpp ran4, ff wg2


18/02/2010

7

OFDMA interference pattern in DownlinkNo intra-cell interference – orthogonal resource blocksInter-cell Interference with perfect synchronizationI t ll I t f ith ti i ff t1Inter-cell Interference with timing offset1

( ) 0< t cpl tδ Δ <⎧⎪ 2

( )

sin( ( )/ )( , ) < t

0 2

cp

cpcp cp

l t t TI t l t T t

lT t t T t

ππ

⎧⎪

−Δ⎪Δ = Δ < +⎨⎪⎪ + <Δ < +⎩0 2cp cpT t t T t+ <Δ < +⎩

if is uniformly distributedtΔ

2 3 0

6( )( )

1 0

cp

cp

T tl

T tI l

T l

+⎧=⎪ +⎪= ⎨

⎪2

1 02( ) ( )cp

T lT t lπ

⎪ ≠⎪ +⎩

1 – J-W. Lee, ”Other-Cell Interference Reducing Resource Allocation in OFDM-Based Asynchronous Cellular Systems”, EURASIP Journal on Wireless Comm. and Networking, 2008


18/02/2010

8

Interference Types


18/02/2010

9

How serious are they?Evaluate the influence of different types of interferences

mainly on Link budget study Computer simulationsComputer simulations Field measurementsResults1Interference Types Influence Level Note

Femto-to-Femto Interference in Femto uplink

Moderate Rare to occurTx power limit for Femto UE

Femto-to-Femto Interference in F t d li k

Moderate Adaptive power controlP t d dFemto downlink Prevent dead zones

Macro-to-Femto Interference in Femto downlink

Low Num. of affected users small

Femto to Macro Interference in Moderate potentially Limit Femto UE powerFemto-to-Macro Interference in Macro uplink

Moderate, potentially high

Limit Femto UE power

Macro-to-Femto Interference in Femto uplink

Low Near-far effect

Femto-to-Macro Interference in Macro downlink

High Handover, adaptive Femto BS power control

1 – “Interference Management in OFDMA Femtocells”, Femtoforum WG2 whiter paper, 2009


18/02/2010

10

Power control algorithmsPC algorithms existing in LTE macro cell

Downlinkbasically constanty

UplinkOpen loop – Fractional power controlClosed loop – link SINR measurement and control signaling via DLp g g

PC algorithms for LTE femto cell to developDownlink

t t t itnon-constant transmit power

UplinkOpen loop – modified Fractional power controlCl d l h ti l ll tiClosed loop – may approach optimal power allocation

Advanced PC algorithmsJoint PC with sub-band allocation etc. – additional degrees of freedomCooperative PCCooperative PC


18/02/2010

11

Algorithms for Downlink femto cell PCto mitigate the Femto-to-Macro interference in Downlink1 Based on distance to closet1. Based on distance to closet Macro eNB

Femto eNB’s location is knownfar distance = reduce HeNB power

2 Based on pathloss to closet2. Based on pathloss to closet Macro eNB

measure signal strength of Macro DLg g

3. On detection of victim UEsl t t h d donly protect when needed

if detected, strict HeNB power controlif not detected, soft power control


p18/02/2010

12

Downlink PC with Connectivity issuesill “cell breathing” raises interference


18/02/2010

13

Further on “cell breathing”

Femtocell expands to unburden “Hot Spot” load

Femtocell shrinks in “Cold Spot”


18/02/2010

14

Algorithms for Uplink femto cell PC – Open LoopConventional Fractional power control

max minmin 1, max ,tPLP P R

γ⎧ ⎫⎡ ⎤⎛ ⎞⎪ ⎪⎢ ⎥= × ⎨ ⎬⎜ ⎟⎢ ⎥

is the power cap (cell specific)

max min, ,tx ilePL −

⎢ ⎥⎨ ⎬⎜ ⎟⎢ ⎥⎝ ⎠⎪ ⎪⎣ ⎦⎩ ⎭

maxPis the fractional factor

Modifications on FPCMacro eNBs adjust of the Femtocells in their vicinity

γ

PMacro eNBs adjust of the Femtocells in their vicinityhigher UL interference -> reduce

Macro eNBs adjust or of the Femtocells in their vicinityis a balancing factor between throughput and system coexistence

maxPmaxP

γγ x ilePL −

is a balancing factor between throughput and system coexistence

Parameter sets proposed by 3GPP RAN41

γ

Parameter set Gamma (γ) PLx-ile

10 MHz bandwidth

5 MHz bandwidth

Set 1 1 112 115

Set 2 0,8 129 1331 – 3GPP R4-091796, May 2009


18/02/2010

15

Algorithms for Uplink femto cell PC – Closed Loopoptimal power controls

interference limited network, e.g. CDMAassign target SIRassign target SIRmeasure current UL SIRfeedback to UE via DL

minimize

subject to ( ) , for

ii

i i

p

SIR iγ≥ ∀

∑pj ( ) ,

variables i iγp

p

γ[ 1] [ ][ ]

ii i

i

p t p tSIR tγ

+ =


18/02/2010

16

Fractional Frequency Reuse with PCTraffic demand on Femtocell is relatively lowOptimal FFR allocation with joint power controlRandomly FFR allocationa do y a oca o


18/02/2010

17

Joint PC and FFRFFR scheme 1 (baseline) FFR scheme 2


18/02/2010

18

Joint PC and FFRFFR scheme 3 FFR scheme 4


18/02/2010

19

Preliminary results 500x Monto Carlo simulation

30

35sum. TX Power vs. Num of Femtos

FFR3 FFR4 FFR1 FFR2

25

Pow

er (d

Bm

)

15

20

aver

age

Sum

TX

P

10

0 50 100 150 200 250 300 350 400 450 5005

Num of Femto BSs


18/02/2010

20

Preliminary results 500x Monto Carlo simulation

4.5

5Outage Probability vs. Num. of Femtos

FFR3 FFR4 FFR1 FFR2

3

3.5

4

abili

ty (%

)

FFR3 FFR4 FFR1 FFR2

1 5

2

2.5

Pow

er O

utag

e P

roba

0.5

1

1.5

0 50 100 150 200 250 300 350 400 450 5000

Num. of Femto BSs


18/02/2010

21

Besides FFR and power control Smart scheduling

Femtocell and Macrocell cooperative schedulingInterference sensing and opportunistic schedulingInterference sensing and opportunistic schedulingTime, frequency, antenna degrees of freedom

Operators indoor band sharingWin-win to all

TDD at UL FDDU li k h l i FDD i t l d dUplink channel in FDD is not always crowdedTDD works well for indoor


18/02/2010

22

Concluding remarksFemto cell is a promising technology

I t f i f th j bl f f t llInterference is one of the major problems for femto cell

Power control can mitigate some interference but not allPower control can mitigate some interference but not all

Advanced interference mitigations have to be developedAdvanced interference mitigations have to be developed

Lots of works need to be done in experiment and measurementLots of works need to be done in experiment and measurement


18/02/2010

23

Thank you for your attention

Documents

Power control for interference management in LTE femto ... PC for interference mgmt in LTE Femto Zhao Zhao, Thomas Kaiser 18/02/2010 2. Why Femto at home? How many mobile phone calls