interference management in lte femtocell systems using fractional frequency reuse

7/29/2019 Interference Management in LTE Femtocell Systems Using Fractional Frequency Reuse

1/5

Interference Management in LTE Femtocell SystemsUsing Fractional Frequency Reuse

Poongup Lee, Taeyoung Lee, Jangkeun Jeong, and Jitae ShinSchool ofInformation andCommunication EngineeringSungkyunkwan University, Suwon, 440-746, Korea{poongup, kingcarlee, carrot252, jtshin}@skku.edu

Abstract- Recently, Long Term Evolution (LTE) has developeda femtocell for indoor coverage extension. However,interference problem between the femtocell and the macrocellshould be solved in advance. In this paper, we propose aninterference management scheme in the LTE femtocell systemsusing Fractional Frequency Reuse (FFR). Under the macrocellallocating frequency band by the FFR, the femtocell choosessub-bands which are not used in the macrocell sub-area toavoid interference. Simulation results show that proposedscheme enhances total/edge throughputs and reduces the outageprobability in overall network, especially for the cell edge users.Keywords- Femtocell, Interference management, FractionalFrequency Reuse (FFR), Long Term Evolution (LTE).

I. INTRODUCTIONLong Term Evolution (LTE) is one of the most promisingcandidates for next generation communication standard. TheLTE is technologically based on Orthogonal FrequencyDivision Multiple Access (OFDMA) to achieve higher datarates and enhanced spectral efficiency. The frequency andtime resources are allocated to users in orthogonal manner. Akey feature is Frequency Reuse Factor (FRF) of 1 formaximal resource use. However, if same sub-carriers areused by different users among adjacent cells, Co-ChannelInterference (CCI) problem occurs especially for cell edgeusers. Appropriate inter-cell interference coordinationtechnique should be required to enhance the system capacity[1].Recently, the LTE has developed a femtocell for indoorcoverage extension. The femtocell is defmed as very smallsize, low-power home base station that works in the licensedfrequency bands, and it is connected to broadband Internetbackhaul [2]. The femtocell brings various benefits to bothconsumers and operators, such as enlarged indoor coverage,enhanced system capacity, Quality of Service (QoS), and

~ e d u c e d capital and operation expense. Due to increasingmdoor phone calls and data services but insufficientmacrocell coverage, the femtocell could be an attractivesolution.However, interference problem between the femtocell andthe macrocell should be solved in advance, because the

femtocell is deployed over the existing macro network, and ituses same spectrum with the macrocell. Dedicated channelapproach is one of the easiest ways to solve this problem, butthe frequency resources are not utilized effectively. Co-channelmodel is suitable for practical deployment, but the CCIproblem should be solved [3]. The femtocell is often turned onand off, and installed at unknown location, because it ismanaged by a personal customer, not by a network operator.Therefore, many parameters should be automatically adjustedto avoid the CCI.Fractional Frequency Reuse (FFR) is discussed in theOFDMA based network such as the LTE, to overcome the CCIproblems [4-6]. In the FFR, whole frequency band is dividedi n t ~ several sub-bands, and each sub-band is differentlyassigned to center zone and edge region of the cell. While reusefactor of the center zone is one, the edge region adopts biggerreuse factor. As a result, intra-cell interference is removed, andinter-cell interference is substantially reduced. At the same timethe system throughput is enhanced.In this paper, we propose an interference managementscheme in the LTE femtocell systems using the FFR. Under themacrocell allocates frequency band using the FFR, thefemtocell chooses sub-bands which are not used in themacrocell sub-area to avoid interference.The remaining part of this paper is organized as follows.Related work is explained in Section II. Section III describesthe proposed interference management scheme. The simulationresults are shown in section IV. Finally, we conclude our paperin section V.

II . RELATED WORKInterference management issues for the femtocell systemshave been actively discussed both in the LTE and in theWorldwide Interoperability for Microwave Access (WiMAX)[6].Spectrum partitioning and Frequency ALOHA (F-ALOHA)scheme was proposed [7]. It removes interference betweenmacrocells and femtocells by assigning orthogonal spectrum.Among the femtocells, each femtocell uses sub-channels inrandom manner. Optimal spectrum portion for the femtocellscan be determined. However, this scheme is based on the

ISBN 978-89-5519-146-2 - 1047 - Feb. 7-10, 2010 ICACT2010


2/5

dedicated channel approach. Although the spectrum isutilized adaptively, the full bands are not serviced for themacrocell.Dynamic Frequency Planning (DFP) algorithm wasanother proposal for interference avoidance [8]. Afterdividing the OFDMA network into several sectors, theamount of sub-channels is estimated considering userbandwidth demand in each sector. Interference amongsectors are calculated when the sectors transmit the samefrequency. Optimization function is run to minimize theoverall network interference. However, this scheme follows acentralized network structure, which is not appropriate forfemtocell managed by a personal owner.Several heuristic algorithms for frequency assignmentwere investigated [9]. The recommended one is LeastInterference Power (LIP) algorithm, where a powered-upfemto base station chooses a frequency segment thatminimizes the interference level. In turn-on orders algorithm[9], frequency allocation is conducted according to the orderof the femtocell turned on. However, the femtocells aredeployed in the form of rectangular matrix, not in the randommanner. Also, interference between the macrocell and thefemtocell is not evaluated.Isolated and coupled model considering user location wasdiscussed for OFDMA femtocells [10]. In the isolatedmodel,macro and femto users are allocated different resources splitby time and frequency slots. The coupled model reusesmacro resources for femtocells which are located in the cellboundary, while the center zone femtocells use orthogonalresources from the macrocell. However, these schemes donot apply the FFR concept for the OFDMA system.Frequency reuse and pilot sensing scheme was proposedto reduce the CCI [11]. After applying the FRF 00 or aboveto the macrocells, the femtocells use the remaining frequencysub-bands. For example, if the reuse factor is three and themacrocell uses sub-band I among the three sub-bands I, II,and III, the femtocell chooses sub-band II and III. However,the macrocell throughput is reduced, even though the overallcapacity is enhanced. Also the reuse factor in macrocell isagainst the OFDMA based network such as the LTE and theWiMAX,where the target of the reuse factor is one.The FFR is one of the solutions to reduce inter-cellinterference in macrocell system, especially for the cell edgeusers. Also it is helpful to achieve the reuse factor of one.Under this condition, the interference from the femtocelldeployment should be minimized for the macro users. So, wefocus on the interference management between the macrocelland the femtocell using the FFR.

III.PROPOSED SCHEME IN FEMTOCELLA. IllustrationoftheProposedSchemeWe allocate the frequency sub-bands into macrocell andfemtocell as depicted in Figure I. The macrocell coverage isdivided into center zone and edge region including threesectors per each, denoted by Cl, C2, C3, andEI , E2, E3. Thewhole frequency band is partitioned into two parts, and one

part of them is classified into three pieces again, totally denotedby A, B, C, and D.For macrocell, different frequency sub-band is allocated tothe each macrocell sub-area according to the FFR. The reusefactor of one is applied in the center zone, while the edgeregion adopts the factor of three. The sub-band A is used in thecenter zone (Cl, C2, and C3), and sub-band B, C, and D isapplied in the El , E2, and E3 regions, respectively.Under this circumstance, the femtocell chooses sub-bandswhich are not used in the macrocell sub-area. Especially whenthe femtocell is located in the center zone, the femtocelladditionally excludes a sub-band which is used by macrocell inthe edge region of current sector. For example, when afemtocell is located in region El , it uses sub-band A, C, or D,while the macrocell uses sub-band B. If a femtocell ispositioned in zone C1, sub-band C and D is applied. Thefemtocell avoids sub-band A which is used by macrocell inzone Cl . Also it avoids sub-band B which is used by macrocellin region El , because the received signal power of sub-band Bis relatively strong for that femtocell.Due to the characteristics of the OFDMA, the macrocell isinterfered by inter-cell, and that interference is furthermitigated by the FFR. The femtocell uses different sub-band toavoid interference from the macrocell. The sub-band is reusedin the macrocell coverage as much as possible, becausetransmit power of the femtocell is very small. Therefore, theinterference between the macrocell and femtocell is greatlyavoided. Also more sub-carriers are allocated to femtocellwhich is located in the edge region, in order to improve theperformance of the edge users.uMacro: BFemIa: A,C,D

BCD

Figure I. The proposed interference management scheme using FFR

B. PerformanceFormulationWe formulate downlink Signal to Interference and NoiseRatio (SINR) and system throughput. The overall network iscomposed of two-tier 19 macrocells, and many femtocells arerandomly deployed over the macrocells. A macro user isinterfered from neighboring 18 macrocells and all of theadjacent femtocells. Due to small transmit power, onlyfemtocells which are located in the l-tier macrocell area giveinterference to macro user. The received SINR of a macro user

m on sub-carrier k can be expressed as [4]

ISBN 978-89-5519-146-2 - 1048 - Feb. 7-10, 2010 ICACT 2010


3/5

where, {3m,k notifies the sub-carrier assignment for macrousers. When {3m,k = 1, the sub-carrier k is assigned to macrouser m. Otherwise, {3m,k = O. From the characteristics of theOFDMA system, each sub-carrier is allocated only onemacro user in a macrocell in every time slot. This impliesthat L ::1 f3m,k = 1, where Nm is the number ofmacro users ina macrocell.Similar expression for femto users related to the practicalcapacity and the overall throughput is possible except" ,N f f3 = 3 .Nfis the number offemto users in a macrocell.L..J/=1 /, kThis implies that the proposed scheme reuses the fullfrequency band three times in a macrocell.C. OperationalAlgorithmThe macro users are allocated the sub-carriers accordingto the FFR as shown in Figure 1. The center zone and the

P GSINR = M,k m,M,k (1)m,k Noi1f + L PM ',kGm,M',k + L PF,kGm,F,k

M' Fwhere, PM,k and PM',k is transmit power of serving macrocellM and neighboring macrocell M' on sub-carrier k,respectively. Gm,M,k is channel gain between macro user mand serving macrocell M on sub-carrier k.Channel gain fromneighboring macrocells are denoted by Gm,M',k. Similarly, PF,kis transmit power ofneighboring femtocell F on sub-carrier k.Gm,F,k is channel gain between macro user m and neighboringfemtocell F on sub-carrier k.No is white noise power spectraldensity, and ~ f i s sub-carrier spacing.In case of a femto user, it is interfered from all 19macrocells and adjacent femtocells. The received SINR of afemto userf on sub-carrier k can be similarly given by

P GSINR = F,k j,F,k (2)j,k Noi1j + LPM,kGj,M,k +LPF',kGj,F',k

M F'The channel gain G is dominantly affected by path loss,which is different for outdoor and indoor. The path loss foroutdoor is modeled as PLoutdoor = 28+3510g10(d) dB, where dis the distance from a base station to a user in meters.Otherwise, indoor model is PL;ndoor = 38.5+2010g10(d)+LwallsdB, where Lwalls is 7, 10, and 15 dB for light internal, internal,and external walls, respectively [3]. So, the channel gain Gcan be expressed as

G =10-PL ll o (3)The practical capacity of macro user m on sub-carrier kcan be given by [4]

Cm,k =i1jlog2(1+aSINRm,k) (4)where, a is a constant for target Bit Error Rate (BER), anddefmed by a = -1.5/ln(5BER). Here, we set BER to 10-6The overall throughput of serving macrocell M can beexpressed as

(6)

t: = ~ P CM L..J L..J m,k m,km k

(5)

three sectored edge region use the separated frequency subband.Under this condition, a femtocell senses the neighboringmacrocell signals, when it turns on. The Received SignalStrength Indication (RSSI) values are compared for each subband A, B, C, and D. When the RSSI of sub-band A is strong,the femtocell is located in the center zone. In addition, if thesub-band B signal is strong, the femtocell location is zone C1.The femtocell chooses sub-carriers from the sub-band C and D.It excludes the sub-band A and B, whose signal power is strongbecause these are used by the macro users. It is similar for zoneC2 and C3.On the other hand, if the RSSI of sub-band A is weak, thefemtocell is positioned in the edge region. Moreover, if the subband B signal is strong, the femtocell position is region E1. Thesub-carriers from the sub-band A, C, and D are selected by thefemtocell. The sub-band B is excluded because its RSSI is high.Similar scheme is possible for the region E2 and E3.IV. SIMULATION RESULTS

A. Scenarios and EnvironmentsWe evaluate the proposed scheme in terms of throughputand outage probability using SINR threshold in the range of 0to 30 dB. We also concentrate on the performance of the celledge users. The outage is determined when the SINR level of asub-carrier do not exceed the designated threshold. The outageprobability Pout is given byLu L k Ju,k SINRu,k

Pout =" " f3. SINRL..J uL..Jk u,k u,kwhere, Ju,k indicates failed sub-carrier assignment for user u onsub-carrier k. If Ju,k =1,the SINR of that sub-carrier is under theSINR threshold (SINRu,k < SINRthreshold). So, the ratio betweenthe number of sub-carriers under the SINR threshold and thenumber of total sub-carriers is the outage probability.The proposed scheme is compared with several comparisonschemes as follows. In FFR-3 scheme, FFR is applied to themacro users, and femto users are randomly assigned from fullfrequency band. On the other hand, FFR is not adopted formacro users inNoFFR-3 scheme. The amount of total availablesub-carriers for femto users is three times of the full band, asequal to the proposed scheme. These features are summarizedin Table 1. The Amount column implies the amount of usedsub-carriers denoted by the sub-carrier assignment parameter {3.Table 1. Summary ofProposed and Comparison Schemes

Schemes Macro user Femto userFrequency Amount Frequency AmountDivide centerProposed FFR and edges

1 (Figure 1) 3FFR-3 FFR RandomNoFFR-3 Random RandomNote: Amount column implies value of I::l Pm,k and I ; ~ l Pj,kfor macro and femto users, respectively.

ISBN 978-89-5519-146-2 - 1049 - Feb. 7-10, 2010 ICACT2010


4/5

Throughput of Macro+Femto Users (Edge)

Throughput of Macro Users (Center+Edge)

160

150

150

90 120Number of Femtocells

90 120Number of Femtocells

60

60

33:-0- - ----:: ' : -- - ----::':- - - -:-':-::-- - ----: ':-::-- - ---,-'80

1 l , - - ~ - - - ~ - - _ _ _ , _ J , _ = _ _ - : : o : : : : = = = C J30Figure3. Throughput ofmacro and femto users located onlyin the edgeregionas the number offemtocells varies

Figure 2. Throughput ofmacro users located in center zoneand edgeregionas the number offemtocells varies

I- - - - - - - - - - ~ - - - - " ' B - - - - - 6 l - - - - ]I I I I9 - - - - - ~ - - - - - - ~ - - - - - - 1 - - - - - - 4 - - - - - -I I I II I I I-- ~ - - - - - - T - - - - - - I - - - - - - ~ - - - - - -I I I I

I I I- - - - - T - - - - - - I - - - - - - ~ - - - - - -I I I I

6 - - - - - ~ - - - - } - - - - - - : - - - - - - ~ - - - - - -I I I5 l _ _ _ _ _ _ _ __ J _I II Y ~ = : l H

4 L _II

between the macro user and the femto user is higher than theproposed scheme.The gap of throughput between the proposed and the otherschemes is bigger when more femtocells are deployed. Themacro user is interfered more from the femtocells as thenumber of femtocells increases. However, the proposed schemeavoids this interference and it shows less degradation.In Figure 3, it describes the throughput of total users locatedat the edge region only. In the OFDMA cellular network, theperformance of the edge region is poor due to the inter-cellinterference. In the proposed scheme, allocating the more subcarriers to femtocells which are located in the edge region, thethroughput of the edge region is improved.Otherwise, compared schemes assign the sub-carriersirrespective of the user location which is center zone or edgeregion. The throughput is smaller than the proposed scheme inthe edge region. Also, FFR-3 scheme shows higher throughputthan NoFFR-3 scheme, because it reduces interference byapplying FFR to the macro users. The gains of the proposedscheme are 27% and 47% in average, when compared with theFFR-3andNoFFR-3 schemes, respectively.The simulation parameters are listed in Table 2. Theoverall network is composed of two-tier 19 macrocells, andfemtocells are randomly deployed over the macrocells. Allthe base stations are operated by the OFDMA technology.We vary the number of femtocells from 30 to 180 in onemacrocell coverage to change simulation environment. Themacro and femto users are randomly distributed in theoverall network. The macro and femto users are randomlyallocated the available sub-carriers from the designated rangein each scheme. When the FFR is applied to macro user, ahalfof the full band is assigned for center zone, the other foredge region. An onmi-directional antenna and three sectorantennas are installed at a macro base station, for center zonewith transmit power of 15Wand for edge region with 22 W,respectively. On the other hand, if the FFR is not used bymacro user, the transmit power of a macro base station is 20W. All the femtocells use 20 mW as the transmit power. Thedifferent channel model is used for indoor and outdoor,where the path loss is a dominant factor. Then, we find outthe downlink received SINR values for each user and eachsub-carrier. Using this value, the throughput and the outageprobability are calculated via users located in the centralserving macrocell of 19cells.B. Numerical Results andComparisonFigure 2 shows the throughput of macro users located in

the macrocell coverage as the number of femtocells varies. Inproposed scheme, the femto users can get the sub-carrierswhich are not used by macro users at each location. So, theinterference between macro users and femtocells is greatlyavoided. The femtocells affect to macro users less than thecomparison schemes.On the other hand, the sub-carriers are randomly assignedto the femto users regardless of the sub-carriers used by themacro user in the FFR-3 and NoFFR-3 schemes. The samesub-carriers can be used by the macro user and the femtousers who are very close to each other . So, interference

Table 2. Simulation ParametersParameters ValuesMacro Femto

Number of cells 19 (two-tier) 30-180/macroCell coverage 280m 30 mBS transmit power FFR: 15,2 2 W 20mWw/oFFR: 20 WNumber of users in one macro 180 180cell coverageChannel Bandwidth 5 MHzFFTsize 512Number of occupied sub- 300carriersSub-carrier spacing 15 kHzWhite noise power density -174 dBm/HzSize of center zone 0.63 ofmacro coverage

PLoutdoor = 28+3510glQ(d)Channel model PLindor = 38.5+2010glQ(d)+Lwalls(Path loss, PL) Lwalls= 7 dB, if din ( 0,10]Lwalls= 10 dB, if din (l 0,20]Lwalls= 15 dB, if din (20,301

ISBN 978-89-5519-146-2 - 1050 - Feb. 7-10, 2010 ICACT 2010


5/5

Outage Prob. of Macro+Femto Users (Center+Edge)

Figure4. Outage probability of macro and femto users according toSINR threshold varies, when 30 femtocells are deployed

V. CONCLUSIONSIn this paper, we propose an interference managementscheme in the LTE femtocell systems using the FFR. Theproposed scheme enhances throughput and reduces outageprobability, especially for the cell edge users. It is beneficial

for the OFDMA network such as the LTE, where the FFR isapplied.ACKNOWLEDGMENT

This work was supported by the National ResearchFoundation (NRF) grant funded by the Korea government(MEST) (2009-0071289).REFERENCES

[I] O. Boudreau et al., "Interference Coordination and Cancellation for 40Networks," IEEE Commun. Mag., vol. 47, no. 4, pp. 74-81, Apr. 2009.[2] V. Chandrasekhar, J. Andrews, "Femtocell Netwoks: A Survey," IEEECommun. Mag., vol. 46, no. 9, pp. 59-67, Sept. 2008.[3] L. Ho, H. Claussen, "Effects of User-Deployed, Co-Channel Femtocellson the Call Drop Probability in a Residential Scenario," IEEEInternational Symposium on Personal , Indoor and Mobile RadioCommunications (PIMRC),Sept. 2007.[4] H. Lei, L. Zhang, X.Zhang, and D. Yang, "ANovel Multi-cell OFDMASystem Structure Using Fractional Frequency Reuse," IEEEInternational Symposium on Personal , Indoor and Mobile RadioCommunications (PIMRC),Sept. 2007.[5] M.Assaad, "Optimal Fractional Frequency Reuse (FFR) inMulticellularOFDMASystem," IEEE Vehicular Technology Conference (VTC), Sept.2008.[6] S. Yeh et al., "WiMAX Femtocells: A Perspective on NetworkArchitecture, Capacity, and Coverage," IEEE Commun. Mag., vol. 46,no. 10, pp. 58-65, Oct. 2008.[7] V. Chandrasekhar, J. Andrews, "Spectrum Allocation in Two-TierNetworks," IEEE Asilomar Conference on Signals, Systems andComputers,Oct. 2008.[8] D. Lopez-Perez et al., "InterferenceAvoidance and Dynamic FrequencyPlanning for WiMAX Femtocells Networks," IEEE InternationalConference on Communication Systems (ICCS),Nov. 2008.[9] H. Zeng, C. Zhu, and W.Chen, "System Performaceof Self-OrganizingNetwork Algorithm in WiMAX Femtocells," ACM InternationalConference on WirelessInternet (WICON),Nov. 2008.[10] K. Sundaresan, S. Rangarajan, "Efficient Resource Management inOFDMA Femto Cells," ACM International Symposium on Mobile AdHoc Networking and Computing (MobiHoc),May 2009.[11] T. Kim, T . Lee, "Throughput Enhancement of Macro and FemtoNetworks By Frequency Reuse and Pilot Sensing," IEEE InternationalPerformance, Computing and Communications Conference (IPCCC),Dec. 2008.

3050 15 20SINRThreshold (dB)

5

II I I I0.9 - - - - ~ - - - - - r - - - - ~ - - - - - r - - - -I I I I0.8 LI I II IQ7 - - - - ~ - - - - - ~ - - -I II I----1-----I I____ ...I L __

I

0.62lif. 0.5s.s 0.4o

Figure 4 depicts the outage probability of total usersaccording to the SINR threshold, when 30 femtocells aredeployed in the macrocell coverage. In a given SINRthreshold, the proposed scheme indicates lower outageprobability than other schemes. Also, proposed schemedecreases the outage probability more at low SINR threshold.This implies that the proposed scheme effectively supportsmore users, even though the interference is severe.As shown in these results, our proposed scheme enhancesthe overall throughput and reduces the outage probability,especially for the cell edge users.

ISBN 978-89-5519-146-2 - 1051 - Feb. 7-10, 2010 ICACT 2010

interference management in lte femtocell systems using fractional frequency reuse

Documents