Through the LTE enhance relay network performance--澳洲Paper代写范文

澳洲Paper代写范文:“Through the LTE enhance relay network performance”，这篇论文主要描述的是通过利用LTE技术来满足IMT-Advanced的框架要求，进而达到继电保护的效果，这是一门成本较低且有远大发展前景的技术，在这一项技术的支持下能够将下行峰值数据和上行数据的带宽得到显著的提高，性能提高的同时也能够降低信号的干扰噪声比，能够满足市场用户日益增长的用户需求。

Abstract— Relaying is considered a promising cost-efficient solution in 3GPP LTE-Advanced for coverage extension and throughput enhancement. Due to compact physical characteristics and low power requirements of the relay nodes, the relay deployment has a high degree of freedom. In this paper, the impact of site planning on the backhaul performance of relay networks within the LTE-Advanced framework is investigated. The site planning strategies are formulated in terms of SNR and SINR based selection criteria. 3GPP compliant simulations are performed comparing relay site planning strategies. It is shown that proper site planning yields significant SINR gain on the eNB-RN link along with a clear reduction in the shadowing standard deviation compared to random deployment.

I. INTRODUCTION 介绍

Long Term Evolution-Advanced (LTE-Advanced) is the candidate technology of the 3rd Generation Partnership Project (3GPP) which defines the framework for further advancement in LTE to fulfill the requirements of IMT-Advanced (International Mobile Telecommunications Advanced) specified by ITU (International Telecommunication Union). In accordance with these requirements, LTE-Advanced should support peak data rates of 1 Gbps in downlink (DL) and 500 Mbps in uplink (UL), bandwidth scalability up to 100 MHz, increased spectral efficiency up to 15 bps/Hz in UL and 30 bps/Hz in DL, along with improved cell edge capacity, as well as decreased user and control plane latencies [1]. In order to meet these requirements, problems such as low signalto-interference-plus-noise-ratio (SINR) at the cell edge and coverage holes due to shadowing and non-line-of-sight (NLOS) connections should be tackled. Deploying decode and forward relay nodes (RNs) is a promising solution which is one of the proposed technologies for LTE-Advanced networks to meet the growing demand and challenging requirements for coverage extension and capacity enhancement [2][3]. RNs are relatively small nodes with low power consumption, which connect to the core network with wireless backhaul through an evolved Node B (eNB). This feature enables deployment flexibility and eliminates the high costs of a fixed backhaul. Furthermore, RNs do not have strict installation guidelines with respect to radiation, visual disturbance and planning regulation. Therefore, installing RNs involves lower operational expenditure (OPEX) [4] and faster Jyri H.m.l.inen Aalto University, School of Technology and Science P.O. Box 3000, FIN-02015, Finland jyri.hamalainen@tkk.fi network upgrade when operators aim to improve quality of service (QoS) [5]. Thanks to compact physical characteristics and low power consumption of relays, RNs can be mounted on structures like lamp posts with power supply facility. Cell planning and site selection tools are used routinely by operators to improve the system performance and to provide a satisfactory service with minimal deployment expenditure. In this manner, the deployment flexibility of RNs can be utilized by the operators to enhance the system performance by means of improving the eNB-RN link. This becomes even more pronounced in case of interference limited scenarios, where the RN locations exhibiting higher shadowing towards the dominant interferers can be exploited by site selection strategies. This selection results in a better SINR performance and reduces the impact of shadowing. This paper considers two relay site planning strategies within the LTE-Advanced framework, namely the RN cell selection and the RN site location selection. The aim of these planning strategies is to introduce a simple and practical means that can be used to improve the backhaul link between eNB and RN. Both of these strategies can be analyzed by using network planning tools provided that a realistic modeling of path loss and shadow fading is applied. This discussion has been recently carried out in 3GPP standardization. Consequently, a certain bonus has been added to the channel model that describes the backhaul link between eNB and RN, which accounts for the improvement due to relay site planning [6]-[8]. In this paper, we formulate the different site planning strategies using the selection criteria based either on signal-to-noise-ratio (SNR) or SINR. 3GPP compliant simulations are then performed in order to evaluate the effects of different strategies on the system performance. The remainder of the paper is organized as follows. In Section II, the system model including the relay site planning strategies and the simulation assumptions is presented. The simulation results as well as the system performance evaluation for one-tier relay deployment will be provided in Section III. Finally, Section IV concludes the paper.

II. SYSTEM MODEL系统模型

In this section, the relay site planning strategies are analyzed in detail. In addition, the system model is presented and the LTE-Advanced compliant simulation framework is introduced. 978-1-4244-2519-8/10/$26.00 .2010 IEEE 

Figure 1. Different RN deployment scenarios. The eNB on the left is the donor, if the selection is done according to the distance, and the other eNB is typically the interferer. The shadow fading is visualized by the houses. A. Relay Site Planning Strategies In the following, the basis for different site planning strategies is explained using Fig. 1. In this figure, the RN is typically served by the closer eNB (eNB1) and the other eNB (eNB2) is typically the interfering eNB. Some of the backhaul links from the eNBs to the relay positions RN1-RN4 are impacted by shadow fading due to obstacles which are illustrated by the buildings. By selecting a proper RN location, the performance of the backhaul link can be enhanced compared to the case, where a random deployment is used. As can be seen in Fig. 1, both RN1 and RN4 are favorable locations with respect to eNB1 as they experience a better signal quality from eNB1. On the other hand, in order to optimize the SINR, it is also possible to select RN locations that exhibit a higher shadowing towards the dominant interferer like RN3 and RN4 in Fig. 1. If the deployment is noise limited, such a strategy doesn’t provide gain. However, for typical cellular deployments interference will have an impact and there is an incentive to select RN sites accordingly. Both of the above-mentioned approaches can be combined in order to select the location with the best SINR. In Fig. 1 this corresponds to RN4, since it experiences a good signal from the serving eNB1 and at the same time it is not severely interfered by the eNB2. Although it is assumed that the eNB1 in Fig. 1 is always the closest eNB for the RNs, due to shadowing there might be a case such that an RN receives stronger signal from the eNB2 as the eNB1 is shadowed. This effect is most pronounced for RN2 because it suffers from the shadowing towards the eNB1 but not towards the eNB2. In such a case, the RN may select eNB2 as the serving eNB. Following the above-explained arguments, two main relay site planning strategies are considered: A) RN cell selection: There are two options: . (A1): RN always connects to the closest eNB regardless of the shadowing. . (A2): RN is allowed to connect to the best eNB around taking also shadowing into account. B) RN site location selection: Options are now: . (B1): There is only one possible location for RN. . (B2): RN site location can be selected out of M alternatives. Figure 2. Example with an initial location and four other candidate locations. There are totally M=5 alternatives for the RN location. As a result, we obtain four different approaches, namely (A1,B1), (A2,B1), (A1,B2) and (A2,B2). Either SNR or SINR can be used as a criterion for the performance measure for each of these approaches. These criteria are of the form: PkSNRm,k = , (1)PLNm,k Pk Lm,k SNRm,kSINRm,k = = , (2) PN + ΣPk ' Lm,k '1+ ΣSNRm,k ' k '≠kk '≠k where P denotes the signal power, PN is the noise power, L is the path-loss including the shadowing, m refers to the m-th RN candidate location, where m is an integer from the set of [1, M] and k refers to the k-th eNB. We note that k=1 corresponds to the reference eNB and m=1 corresponds to the reference RN location initially deployed in the sector as shown in Fig. 2. In addition, it can be stated that compared to SNR based criterion the SINR based selection criterion takes into account not only the benefit of improving the reception from the serving eNB but also the benefit of reducing the interference caused by other eNBs. Following these criteria the above-mentioned approaches can be expressed as: . (A1,B1): Neither cell selection nor RN site location selection. This is the reference case. SNR(A1,B1) = SNR1,1 , SINR(A1,B1) = SINR1,1 . (A2,B1): Cell selection according to the best eNB, no RN site location selection. SNR1, k0 = max{ SNR1, k }, SINR1, k0 = max{ SINR1, k } . (A1,B2): No cell selection but RN site location selection. SNRm0, 1 =max{SNRm,1}, SINRm0,1=max{SINRm,1} . (A2,B2): Both cell selection to the best eNB and RN site location selection. SNRm0, k0 =max{SNRm, k}, SINRm0, k0=max{SINRm,k} Note that, in (A1,B1) and (A2,B1) both SNR and SINR based criteria will lead to the same result in the selection procedure due to the restriction in site location selection.

TABLE I. SIMULATION PARAMETERS Parameters Default Carrier frequency 2 GHz System layout 7 sites -3 sectors/site Number of RNs 1 + 4 candidate locations Inter-site distance (ISD) 3GPP Case 1 (500m) & Case 3 (1732 m) Distance-dependent Path Loss eNB-RN (Backhaul Link, NLOS) dPL 36.3log10125.2 += d : distance in kilometers Noise power -104 dBm Penetration loss 0 dB Shadow fading Log-normal, 6 dB standard deviation Shadowing de-correlation distance 50 m Shadowing correlation 0.5 between sites 1.0 between sectors eNB parameters Antenna height 25 m (above rooftop) Number of antennas 2 tx, 2 rx eNB transmit power 46 dBm eNB maximum antenna gain 14 dBi Noise figure 5 dB Horizontal antenna pattern . ... . ... .. .. .. ...= m dB H AA ,min 12( ) 2 3θ θθ 25= mA dB, o dB 703 =θ Vertical antenna pattern . ... . ... .. .. .. ....= SLAA dB etilt V ,min 12( ) 2 3θ θθθ o dB 103 =θ , o etilt 15=θ SLA=20 dB RN parameters Antenna height 5 m (below rooftop) Number of antennas 2 tx, 2 rx Maximum transmit power 30 dBm Maximum antenna gain 5 dBi Noise figure 7 dB Antenna Pattern Omni-directional B. System Framework It is assumed that the original radio network planning has been done for a single-hop system and RNs are then introduced to improve the network performance. In this scenario, a UE is either connected to an eNB or an RN. If a UE is connected to an RN, then the end-to-end throughput depends on both the capacity of the access link (RN-UE) and the capacity of the backhaul link (eNB-RN). Hence, the backhaul link capacity is decisive for the resultant system performance. A Matlab based system-level simulator is used for the performance evaluation considering 3GPP Case 1 (ISD 500 m) and Case 3 (ISD 1732 m) scenarios. The simulator follows the Cell Cell Antennabore-sightInitialRN locationsCell Cell Antenna bore-sight Initial RN locations1 2 3 InteInteIntermrmrmediediediaaate tete ReReRegggiiion ononEEEdge dgedge CeCeCentententer rr Figure 3: Different deployment scenarios for RN. The initial location is selected first from the cell edge, second from the intermediate region, and last from the cell center. current 3GPP LTE-Advanced evaluation guidelines of [6] 1 . Since RNs are deployed outdoor, no penetration loss is considered in the path loss. The total path loss in dB scale is then determined by the addition of the distance dependent path loss with a shadowing component which is characterized by a Gaussian distribution with zero mean and 6 dB standard deviation. The shadowing correlation between candidate locations (see Fig. 2) decreases with an exponential rate determined by the de-correlation distance. The normalized auto correlation function is expressed as follows [6][9][10]: Δx. corR()x , (3)Δ=e d where |Δx| is the distance between candidate locations and dcor is the de-correlation distance. We note that the same correlation model has been used in IEEE 802.16j relay evaluation [11]. Detailed simulation parameters are listed in Table I. As the purpose of this work is to analyze the effects of the site planning on the backhaul link, the access link is not taken into account. The initial RN location (midmost location in Fig. 2) can be at the cell edge, on the central cell area or in between as shown in Fig. 3. Four other candidate locations for the RN (m= [2, 3, 4, 5] in (1), (2) and in selection approaches) are defined such that the distance between initial location and a candidate location is 50 m. It is worth to emphasize that the simulated set-up has been used in 3GPP investigations [7][8].

III. PERFORMANCE EVALUATION 绩效评估

We have considered three possible deployments of RNs, namely deploying relays at the cell edge, closer to the cell center and in between these two extremes (intermediate region). These regions are presented in Fig. 3. Both ISDs of 500 m and 1732 m have been investigated. In all cases a single RN is initially located in the bore-sight of the macrocell. 1 Available parameters in the course of this study. The discussion on simulation assumptions is ongoing in 3GPP. 

Figure 4: Gains of different approaches over the reference case (A1, B1). SINR gain at 50%-ile CDF vs. RN deployment scenario is shown for 3GPP Case 1 (ISD 500 m). A. Performance evaluation in 3GPP Case 1 (ISD 500 m) The performances of different approaches are analyzed compared to the reference case (A1,B1). The achieved SINR gains at 50%-ile cumulative distribution function (CDF) depending on the RN deployment region are presented in Fig. 4. The comparison of the selection criteria is done in terms of achievable increase in the system performance. 1) RN deployment at cell edge In this case, the initial RN location is 300 m away from the macrocell site center and about 30 m away from the cell edge. According to Fig. 4, the SINR based selection criterion outperforms the SNR based criterion by about 2 dB. As expected (A2,B2) achieves the best gain for the RN deployment at the cell edge. On the other hand, for (A2,B1) both SNR and SINR based criteria yield the same result. 2) RN deployment at intermediate region Here the initial RN location is 250 m away from the macrocell site center and about 80 m away from the cell edge. As Fig. 4 illustrates, the SINR based selection criterion is clearly better than the SNR based selection criterion by again about 2 dB. In this region, (A2,B2) and (A1,B2) achieve the same gain, since the closer the RN is to the eNB the higher the probability is that the cell selection results in favor of the closer eNB. Moreover, since the RN is closer to the eNB, the gain obtained from the site planning is smaller than in the previous cell edge deployment. 3) RN deployment at cell center Now the initial RN location is 170 m away from the macrocell site center. As it can be seen in Fig. 4, the difference between the gains obtained using the SINR and SNR selection criteria (about 1.5 dB) is smaller than in the previous cases because relays are closer to the eNB and therefore, the impact of the interference is lower. In addition, the gain obtained from relay site planning becomes lower. In general, results of Fig. 4 show that RN’s ability to connect to the best available eNB is beneficial especially at the cell edge. Such a property can be a part of the RN’s self-Figure 5: Gains of different approaches over the reference case (A1, B1). SINR gain at 50%-ile CDF vs. RN deployment scenario is shown for 3GPP Case 3 (ISD 1732 m). configuration that is needed to make RNs easily deployable and cost-efficient. Moreover, the selection of the RN site location out of a few candidate locations is even more beneficial, since it clearly suppresses the negative impact of the shadowing and the interference from the adjacent eNBs. B. Performance evaluation in 3GPP Case 3 (ISD 1732 m) While 3GPP Case 1 represents a strongly interference limited urban scenario, in the 3GPP Case 3 the impact of interference is smaller due to increased ISD. Also in this scenario the performances of different approaches are analyzed compared to the reference case (A1,B1). The achieved SINR gains at 50%-ile CDF depending on the RN deployment region are presented in Fig. 5. The comparison of the selection criteria is done in terms of achievable increase in the system performance. 1) RN deployment at cell edge The initial RN location is assumed to be 1120 m away from the macrocell site center and about 30 m away from the cell edge. According to Fig. 5, although the SNR based selection criterion performs well, there is still about 2.4-dB higher gain from the SINR based selection criterion. (A2,B2) achieves consistently the best gain. 2) RN deployment at intermediate region When the initial RN location is 1070 m away from the macrocell site center and about 80 m away from the cell edge, the SINR based criterion achieves about 1.8-dB higher gain than the SNR based criterion. Yet, these gains are smaller than at the cell edge. 3) RN deployment at cell center Finally, when the initial RN location is 600 m away from the macrocell site center, it is found that the gains are much smaller than at the cell edge. Besides, the difference between the gains achieved by the SINR and SNR selection criteria are small. Note the similar observation for 3GPP Case 1. It is noticed that in 3GPP Case 3 the gain from RN site planning decays faster than in 3GPP Case 1 when RN is moved from the cell edge towards the cell center. Moreover, 

Figure 6: SINR CDFs for 3GPP Case 1 (ISD 500 m) (a) and for 3GPP Case 3 (ISD 1732 m). (A2, B2) approach is compared to (A1, B1) reference case. In both scenarios the relation of β>αholds, which implies that the shadowing standard deviation reduces due to relay site planning. the gains obtained from the RN site planning are significant at the cell edge, since the shadow fading is a main factor there, which limits the coverage. In addition, the performance difference between SNR and SINR based selection criteria is lower compared to 3GPP Case 1 because of the lower interference. C. RN site planning impact on shadowing standard deviation In the previous section, we showed that the RN site planning benefits in terms of better SINR at 50%-ile CDF. While this is a good measure for the overall performance gain, it is of valuable information to show the whole CDF curve for the best criterion as it is done in Fig. 6. In addition to CDF curves, in Fig. 6 the tangents drawn at the 50%-ile point are shown. The depicted angles of the tangents have the relation of β>α which illustrates the reduction in shadow fading standard deviation due to RN site planning. This reduction amount is about 2 dB for 3GPP Case 1 and about 2.9 dB for 3GPP Case 3. It is important to notice that this decrease in shadowing standard deviation leads to very large gains on areas, where mean path loss is high and the shadowing can be crucial.

IV. CONCLUSIONS 总结

The impact of the RN site planning within the LTE-Advanced framework has been investigated for 3GPP Case 1 (ISD 500 m) and Case 3 (ISD 1732 m). It is shown that remarkable gains on RN backhaul link can be achieved if RNs are a) allowed to make the connection to the best available eNB and b) RN site location can be selected out of a few candidate locations that admit low shadow fading correlation. Two main relay site planning strategies have been introduced based on the experienced SNR and SINR at RN. It is shown that the SINR based selection criterion provides a better approach, whereas especially in 3GPP Case 3 the SNR based criterion becomes comparable. In 3GPP Case 1, the system is interference limited and thus, taking into account the interference from adjacent eNBs is important. It is also emphasized that in practice the RN site planning strategies can be carried out by using the network planning tools provided that shadowing in backhaul links is properly modeled. The performance evaluation of the relay site planning has shown that the SINR gains of 5.2 dB for 3GPP Case 1 and 11.4 dB for 3GPP Case 3 can be achieved when relay nodes are deployed at the cell edge. Furthermore, the standard deviation of the shadowing after the relay site planning can be reduced by 2 dB for 3GPP Case 1 and by 2.9 dB for 3GPP Case 03.

51due留学教育原创版权郑重声明：原创留学生作业代写范文源自编辑创作，未经官方许可，网站谢绝转载。对于侵权行为，未经同意的情况下，51Due有权追究法律责任。

51due为留学生提供最好的服务，亲们可以进入主页了解和获取更多澳洲paper代写范文 提供美国作业代写以及paper辅导服务，详情可以咨询我们的客服QQ:800020041哟。-xz