Many inter-cell robust system throughput and cell edge bit rate performance ... In this paper a Hybrid Inter-Cell Interference the coordination scheme based on ...
A Hybrid Inter-cell Interference Mitigation Scheme for OFDMA based E-UTRA Downlink Jianchi ZHU, Guangyi LIU, Ying WANG, Ping ZHANG WTI Institute, Beijing University of Posts&Telecoms, P.O. Box 92, No. 10, XiTuCheng Road, HaiDian District, Beijing, P.R.China,100876 Email: riwgijcbng163.com Abstract- For the Evolved UTRA (EUTRA), cell edge bit-rate and system throughput become important criteria for the evaluation of the system proposals and requirements on the system performance [1]. Inter-cell interference is the main source of performance degradation for OFDM, especially the data rate of cell edge UEs is severely limited. Many inter-cell interference mitigation schemes are proposed to improve the EUTRA performance, especially the active coordination in the frequency domain resource allocation between neighboring cells [2][3][4]. But they can't achieve robust performance for system throughput and cell edge bit rate when the traffic load varies much. In this paper a Hybrid Inter-Cell Interference Mitigation scheme (HICIM) is proposed, which is a trade-off between system throughput and cell-edge UE performance. From the simulation results when Proportional Fairness (PF) scheduling is adopted in Single Input and Single Output (SISO) environment, the proposed HICIM scheme achieves robust performance of both system throughput and cell-edge UE performance for both high system load and low system load.
Key words- inter-cell interference, E-UTRA, interference mitigation, HICIM, OFDM INTRODUCTION Orthogonal frequency division multiplexing (OFDM) is a transmission technique that has been proposed for wireless high data rate services such as Internet access and multimedia applications due to its ability to overcome frequency selective fading. The proposals for the downlink (OFDMA) of the Evolved UTRA (E-UTRA) support intra-cell orthogonality and as a consequence, the main interference source is intercell interference. The effect of inter-cell interference is particularly detrimental to performance of UEs that are located near the cell edge. Although high system capacity is achieved by OFDMA, cell-edge UE data rate is still too EUTA st sytm tth trafi bae onpce I f b akt slow, Inftre * traffic demand, fortrexample HTTP, T T VoIP, will be the main so improving cell-edge UE performance is important. In order to satisfy a service quality that is largely ineedn of th UE loain it is imotn to conide tehiqe for intrfeec miiato nea th cel edge. So far a lot of techniques have been proposed in order to maximize the throughput and spectrum efficiency in I.
This work is funded byNSF ofChinaundergrantno. 60496312 and 60302024.
I1-4244-0574-2/06/$20.00 (0)2006 IEEE
OFDMA system. In [2], a predefined multiple sub-bands coordination scheme and coordination based on the resource allocation priority are proposed; in [3], the soft frequency reuse scheme for inter-cell interference mitigation is proposed. However, the current proposals can't achieve robust system throughput and cell edge bit rate performance when the traffic load varies much. Based on the current proposals for the inter-cell interference mitigation for E-UTRA, we proposed a Hybrid Inter-Cell Interference Mitigation scheme (HICIM) to achieve robust performance. When the traffic load is low, the coordination scheme based on the resource allocation priority [2] is adopted to mitigate the inter-cell interference; when the traffic load is high, a hybrid scheme of soft frequency reuse [3] and coordination based on the resource allocation priority is adopted. The HICIM is a trade-off between the system throughput and the cell edge bit rate, and it can achieve robust system throughput and cell edge bit rate for both high and low traffic load. For the comparison of different schemes, several schemes of inter-cell interference mitigation are evaluated in this paper. Predefined multiple sub-bands coordination, coordination based on the resource allocation priority and soft frequency reuse scheme are evaluated. For convenient comparison, a resource allocation scheme without inter-cell interference mitigation is demonstrated as a baseline. The organization of this paper is as follows. The main interference mitigation proposals for E-UTRA issued in 3GPP are classified in section II; HICIM is proposed and
analyzed in section III; the Proportional Fair (PF)
scheduling is proposed in section IV; the simulation parameters are given in section V; the simulation results are presented and analyzed in section VI, and the conclusions are drawn in section VII.
II. INTERFERENCE MITIGATION SCHEMES
For the OFDM has no capability to mitigate the inter-cell interferenCe WhiCh iS the main factor to limit the system performance at the cell edge area, the interference mitigation should be taken into account for the performance 2 3 uaranteeing. In the work of 3GPP man roposals p the w o 3GPP, many pos [3] Ig gIAl [4] [5] [6] are issued to mitigate or avoid the inter-cell interference, they can be summarized as three classes: 1. Scheme 1: Predefined multiple sub-bands coordination [2]. A total bandwidth can be segmented into multiple subbands, where neighbor cells or sectors may use the different
sub-bands for the main interference sources in each cell or sector. The number of sub-bands can be as small as 3 when considering 3-sector cell layout as shown in Figure 1, or may be larger than 3 if more detailed UE location-based resources coordination. 3-sector cell structure is adopted in this paper.
3. Scheme 3: Soft frequency reuse scheme for inter-cell interference mitigation [3] In networks with frequency re-use of 1, the downlink throughput at the cell edge is very much limited by interference if interfering terminals in neighbor cells are present. So the interference by the neighbor cell limits the
throughput. In the proposed soft frequency reuse scheme, sub-carriers
are divided into two groups in every cell. One group is called centre sub-carriers group, and the other one is called
Node BI
o
edge sub-carriers group. The centre sub-carriers are used to cover the centre area of the cell, while the edge sub-carriers are used in the edge of the cell, as is showed in figure 3(the C blue represents the inner zones and the other colors Node B3 represent the cell edges). Edge sub-carriers in neighboring A ncell are orthogonal (i.e. no overlap). So, the Soft Frequency ei Reuse (SFRc is the compromise of reuse 1 and reuse 3. _sSubband A1eiubband iubband Although edge sub-carriers in neighboring cell are orthogonal, too many edge sub-carriers will lead to less Frequency resource available in the whole cell and diminish the system Figure 1. Multiple sub-bands coordination throughput. So frequency reuse factor should be a small value while not affecting edge UE performance. To make the Sceme Cordinaion asedon th frequency reuse factor as small as possible, the ratio of the allocation priority [2]. ~~~sub-carriers for the inner and edge area must be optimized Assign different priorities to the sub-bands in the [7]. resource allocation between neighbor cells, which is Here the cell radius is assumed to be R 1Km , and the illustrated in Figure 2. In each cell, the available sub-band is radius of inner area is O. 8R . So the service load of the inner a sliding window and the window size becomes larger as area is: the traffic load higher. Node B2
2. [2: oriainbse.ntersuc 2.lSchemeprort :
resurce
2
IvSubbandsl§iububbandnACl
Node Bi
3Ip , n y _pr
Node B2 Node B3
~
(1)
-(0.8R)2]
(2)
And the service load of the edge area of a cluster is: '2
i
D 77, =-DI(0.8R)2
{D.f[ R 2
So the ratio of C7i and 172 is: (3) 71| 72 (0.8R)2 R2 -(0.8R)2 about osub-carriers are allocated for the inner area, and 60% is allocated to the edge area of the cluster.
pnentI2So
400
=
I0.6
Frequency
Figure 2. Coordination based on the resource allocation priority
In the example of Figure2, the total band is divided into 3 sub-bands to coordinate the interference between 3 neighbor cells. Then, different priorities of resource allocation are assigned to the sub-bands in each cell in such way that the overlapping in the usage of the same sub-band between neighbor cells is minimized. Following the resource allocation rule in Figure 2, especially when the traffic load in each cell is 1/3, resource allocation in frequency domain between 3 cells doesn't overlap, so, the inter-cell interference between 3 cells can be avoided ideally. In this case the frequency reuse factor is 3. When traffic load in each call is 2/3, overlapping of resource allocation in frequency domain between 3 cells is unavoidable. However, it doesn't happen that all the 3 cells use a same sub-band at the same time. In this case the 1 an 3. frequency reuse factor is between
Figure 3 Soft frequency reuse scheme
III. HYBRID INTER-CELL INTERFERENCE MITIGATION SCHEME InscinI,trechmsoitr-llnefrne . .. ' mtgto r rpsd cee1peeie utpesb bands. It separates interference source from neighbor cells into a much larger distance, so the inter- cell interference i S diminished. In scheme 2, different priorities are assigned to
the sub-bands in the resource allocation among neighboring cells, inter-cell interference is avoided especially when traffic load is low. Both of schemes 1 and 2 improve the system throughput promisingly, but the edge UE performance is not so content. Soft frequency reuse is proposed in scheme 3. It improves edge UE performance ultimately but sacrifices the total system throughput. In this paper, we proposed a hybrid inter-cell interference mitigation scheme (HICIM) as a trade-off of the three schemes above. Considering scheme 2 performing ideal inter-cell interference mitigation when the traffic load is low in each cell (i.e.1/3), HICIM adopts this perfect inter-cell interference mitigation when the traffic load is low. When the traffic load is high, coordination based on the resource allocation priority and soft frequency reuse are adopted. When the traffic load is high, in HICIM scheme, first, like soft frequency reuse scheme, sub-carriers are divided into two groups in every cell. One group is called major sub-carriers group, and the other one is called minor subcarriers group. > The major sub-carriers are used to cover the .... whole area of the cell,. while the minor sub-carriers are only used in the edge of the cell. Minor sub-carriers in neighboring cell are orthogonal (i.e. no overlap). Second, like coordination based on the resource allocation priority, different order of priorities is assigned to the major subcarriers in the resource allocation between neighbor cells. In each cell, the available sub-band is a sliding window and the window size becomes larger as the traffic load higher.
=priority A priority B :EMrioity
=p
C
Where R1 (t) is looked up from table 1 [8] with SINR7 . and Tk(t) is the transmission data rate of user k. The sub-carrier i is assigned to user :
l-max IPPrk(t)
(6) L (1-a)Jk(t), else Where 0 < a
