Performance of Repeaters in 3GPP LTE ANTO SIHOMBING

Performance of Repeaters in 3GPP LTE ANTO SIHOMBING Master of Science Thesis Stockholm, Sweden 9

Performance of Repeaters in 3GPP LTE ANTO SIHOMBING Master of Science Thesis performed at the Radio Communication Systems Group, KTH. June 9 Examiner: Professor Ben Slimane

KTH School of Information and Communications Technology (ICT) Radio Communication Systems (RCS) TRITA-ICT-EX-9:67 c Anto Sihombing, June 9 Tryck: Universitetsservice AB

Abstract Repeater communication is one promising candidate solution in future cellular networks because of its ability to increase throughput, data rate and coverage. It is also considered as one candidate technology feature in 3rd Generation Partnership Project (3GPP) Long Term Evolution (LTE) Advanced. Traditionally repeaters have been active continuously and perform blind forwarding without knowing the signal. However the repeater in LTE Advanced is likely to include some advanced functionalities such as frequency selectivity, gain controllability, multi antenna ability, advanced antenna processing, optimum power control algorithm, etc. In this thesis, on-frequency repeaters with frequency selectivity and gain controllability are analyzed and it is shown that the performance of repeater is highly dependent on the environment. It is necessary that the composite path gain (two-hop link) must be better than direct path gain (direct link) and the interference is attenuated in order to use the repeaters. The repeater directional donor antenna can be employed to further improve these two-hop links. And finally the benefit of advanced repeater functionalities is larger for uplink than downlink especially in heavy interference scenario however power limitation is often a bottleneck in uplink. iii

iv

Acknowledgements First of all, I would thank God for His grace and wondrous work in my life and my study. I dedicate this work to my father. I express my gratitude to Peter Moberg and Johan Lundsjö for giving the opportunity to perform this master thesis at Wireless Access Network, Ericsson Research. The experience acquired in the research department, the friendly atmosphere, and the valuable feedbacks have helped me a lot in my studies. I would like also to thank all people in wireless access network department: Aram Antó, Afif Osseiran, Anders Furuskär and Per Skillermark, who gave me essential material and discussions to my thesis work. I would like to thank Prof. Ben Slimane for his help and his availability during this thesis period. The discussion and feedbacks which I had from him are important to my work. The fruitful discussions I had with Bogdan Timus about problem formulation and his previous research in repeater. Thanks to all my friends in wireless system master program who also give encouragements to me and spend over these two years together in KTH. Finally I am also grateful for the moral support and understanding given by my mother, brothers, sister and my fiancée in Indonesia. v

vi

Contents Chapter 1 INTRODUCTION 1 1.1 Background... 1 1. Previous work... 1.3 Problem statement... Chapter LONG TERM EVOLUTION (LTE) 5.1 3GPP LTE Overview... 5. LTE Transmission Schemes... 6..1 OFDM... 6.. SC-FDMA... 7.3 Frame Structure... 9. Scheduling in LTE... 1..1 Downlink Scheduling... 11.. Uplink Scheduling... 1.5 LTE-Advanced... 1 Chapter 3 REPEATER CONCEPT 15 3.1 Repeater Overview... 15 3.1.1 Basic Repeater Design... 16 3.1. Antenna Isolation... 16 3.1.3 On-Frequency and Frequency Shifting Repeaters... 18 3.1. Repeater Delay... 18 3.1.5 Interference and Capacity... 19 3.1.6 Repeater Applications... 3. Two-Hop Communication Model... 1 3.3 Advanced Repeater... 3.3.1 Frequency Selective Repetition... 3 3.3. Repeater Gain Controllability... Chapter SYSTEM MODEL 7.1 General Scenario... 7. Propagation and Channel Model... 7.3 Repeater Model... 9. Simulation Models... 9..1 Radio Network Simulator... 9.. Deployment Scenario... 3 vii

..3 User Generation...3.5 Simulation Parameters...31.6 Performance Evaluation...33.6.1 SINR Calculation...33.6. Throughput...33.6.3 Object Bit Rate (OBR)...33 Chapter 5 SIMULATION RESULTS 35 5.1 Propagation Model...36 5. Number of Repeaters per Cell...38 5.3 Repeater Distance from Base Station...39 5. Repeater Gain... 5.5 Advanced Repeater...3 Chapter 6 CONCLUSION 9 6.1 Conclusion...9 6. Future Work...5 APPENDIX Appendix A Simulation time. 51 Appendix B Propagation Model 5 Appendix C Number of Repeaters per Cell 55 Appendix D Repeater Distance.. 57 Appendix E Repeater Gain 59 Appendix F Advanced Repeater. 61 References 83 viii

List of Tables Table -1 Simulation Parameters... 31 Table 5-1 Performance of repeaters in downlink for different propagation models. 5 Table 5- Performance of repeaters in uplink for different propagation models... 7 ix

x

List of Figures Figure.1 E-UTRAN overall architecture... 6 Figure. A block diagram of SC-FDMA and OFDMA... 8 Figure.3 LTE time-domain frame structure... 9 Figure. LTE subframe and slot structure... 9 Figure.5 LTE frequency-domain structure... 1 Figure.6 Channel dependent scheduling... 11 Figure.7 Downlink resource block assuming normal cyclic prefix... 1 Figure 3.1 Repeater Block Diagram... 16 Figure 3. Antenna Isolation... 17 Figure 3.3 On-frequency (a) and Frequency Shifting (b) Repeaters... 18 Figure 3. A typical repeater s installation in outdoor scenario (a) and indoor scenario (b)... Figure 3.5 Cellular layout of the system... 1 Figure 3.6 Simple illustration of two-hop communication model... Figure 3.7 An illustration of coordinated frequency selective repetition in uplink and downlink... 3 Figure 3.8 Repeaters with different filters... 3 Figure 3.9 Controllable filter banks in the repeater... Figure 3.1 Illustration of frequency selective repeater... Figure 3.11 Example of gain control functionality... 5 Figure.1 Radio Network Simulation... 9 Figure. Repeater deployment illustration... 3 Figure 5.1 Illustration of cellular network deployment with repeaters... 35 Figure 5. A comparison of CDF downlink SINR for different propagation model... 36 Figure 5.3 CDF of uplink SINR for different propagation model... 37 Figure 5. CDF of object bit rate (OBR) for different propagation model... 37 Figure 5.5 Five-percentile downlink object bit rate (OBR) for different number of repeaters per cell... 38 Figure 5.6 Mean uplink object bit rate (OBR) for different number of repeaters per cell... 39 Figure 5.7 Repeater path gain of a moving user for different repeater deployment distance... Figure 5.8 Direct and Composite Path Gain for different downlink repeater gain in WINNER LOS propagation of BS to RN links... 1 Figure 5.9 Mean downlink OBR for different downlink repeater gain... Figure 5.1 Mean downlink OBR for different downlink repeater gain... xi

Figure 5.11 Five-percentile downlink SINR vs mean cell downlink throughput...3 Figure 5.1 Mean downlink SINR vs mean cell downlink throughput...3 Figure 5.13 Five-percentile OBR vs mean cell downlink throughput... Figure 5.1 Mean OBR vs mean cell downlink throughput... Figure 5.15 Five-percentile OBR vs mean cell throughput...7 Figure 5.16 Mean OBR vs mean cell throughput...7 Figure 5.17 Repeater activity in cell radius 166 m and WINNER nlos propagation model scenario...8 xii

List of Abbreviations 3GPP AF AP BS DF FDMA GSM IMT ISI ITU LOS LTE MBSFN MS OBR OFDM OTDOA RN RTT SCM SINR TDMA TTI UTRAN WINNER 3rd Generation Partnership Project Amplify and Forward Access Point Base Station Decode and Forward Frequency Division Multiple Access Global System for Mobile Communications International Mobile Telecommunications Inter-Symbol Interference International Telecommunication Union Line of Sight Long Term Evolution Multi-Media Broadcast over a Single Frequency Network Mobile Station Object Bit Rate Orthogonal Frequency Division Multiplexing Observed Time Difference of Arrival Repeater Node Round Trip Time Spatial Channel Model Signal to Interference plus Noise Ratio Time Division Multiple Access Transmission Time Interval Universal Terrestrial Radio Access Network Wireless World Initiative New Radio xiii

xiv

Chapter 1 INTRODUCTION 1.1 Background In order to ensure the competitiveness of technology, 3rd Generation Partnership Project (3GPP) is considering long term evolution (LTE) as the evolution of third generation (3G) cellular systems. The 3GPP LTE radio-access technology, Evolved Universal Terrestrial Radio Access Network (E-UTRAN), is the future-oriented broadband radio access system whose objective is to obtain higher data rates, low latency, better coverage, improved system capacity, and packet optimized radio access technology []. It will also support end to end services with affordable cost by reducing number of nodes and interfaces while providing enhanced performance and capacity. Meanwhile, the cooperative relay communication for wireless network has recently attracted attention because of its ability to increase the diversity gain in fading environment [1]. The idea behind relay communication model is to improve the cell coverage and capacity in a cost-efficient way. The demand of future cellular networks, especially in densely populated areas, is higher cell capacity than today s system. This implies a denser deployment by adding more base stations (BSs) with a consequence of potential increase in deployment cost. It is shown in [] that the deployment cost of a radio network is proportional to the number of Access Point (AP). Alternatively it is promising to reduce cost by deploying Repeater Node (RN) to substitute base station (BS) while keeping the goal to enhance capacity and coverage. The RN acts like base station but without the need of cable or fiber access, and it uses the same radio access technology from BS to RN and from RN to mobile station (MS). Initial deployment of 3GPP-LTE is expected in 9 and recent news has been announced that a contract of commercial LTE network in Stockholm is signed between TeliaSonera and Ericsson. An initiative has been taken by 3GPP to plan the future work for LTE, referred as LTE-Advanced [6]. LTE-Advanced will be based on LTE, i.e. reuse the main characteristics of LTE, with a selected set of amendments [1] based on International Telecommunication Union (ITU) International Mobile Telecommunications (IMT) Advanced requirements. The context of relaying is also considered as one candidate of technology feature in LTE-Advanced [6]. However, the repeaters considered for LTE-Advanced is likely to include some advanced 1

functionalities such as frequency selectivity, on frequency repetition, network controllability and/or observability. 1. Previous work The base of relaying concept was already established in the 197 s and then it has been proposed as one of the methods to improve capacity in cellular network. Various aspects of transmission are developed in [] where transmission diversity may be employed to improve system performance in term of smaller outage probability. The strategies of diversity transmission include amplifying and forwarding as well as decoding and forwarding. They are also referred as layer-1 (L1) and layer- (L) relays. The deployment strategy of base station (BS) and relay node is also analyzed from a cost perspective [13] [1]. One method was proposed to evaluate the comparison between different deployments which is called isoperformance curve [13] [1]. 1.3 Problem statement Repeater which is used to refer L1 type of relay is studied as the focus area in the thesis. They are considered to be simpler than L/L3 relays and are transparent to the system. Traditionally, repeaters are active all the time and perform "blind" forwarding without knowing the received signals. Repeater in Global System for Mobile communications (GSM) and 3G systems is one example of commercial products which are widely used now [7] [8]. On-frequency repetition, in the other hand, is used at repeaters in order to avoid duplex loss including joint use of signals from the direct link and repeated link. The study of this thesis involves on learning about LTE and LTE-Advanced system, and how repeaters can be incorporated in such networks. In these systems, repeaters have the advanced ability to select which frequencies used to forward the signals with certain gain. If a resource block is assigned to a user, the repeater will amplify and forward the signal in this resource block. The aim of this thesis is to quantify the benefits and gain of advanced repeater in 3GPP LTE. There is no doubt that repeaters can be beneficial in specific scenarios, but the question is how large the gain is and how much the system performance is affected by the increased interference that is an inherent characteristic of a repeater. We will see the effects of different propagation models, number of repeaters per cell, repeater distance and repeater gain toward the cell loads. The performance metrics used for the evaluation are signal to interference and noise ratio (SINR), cell throughput, object bit rate (OBR), and repeater activity. It would be more interesting to look at the 5-percentile of SINR and throughput which correspond to target users and the main objective to be improved in repeater deployment scenario. The deployments of repeater could be in a regular pattern with specific radius from the base station, deployed on cell edge, inside buildings [19] or other alternatives. For simplicity, the deployment inside building is assumed to be the same with outside building, i.e. regular pattern, but it is simulated together with

simple indoor propagation model which is the outdoor propagation model plus wall attenuation loss. Propagation model used in the system is aligned with 3GPP spatial channel model (SCM) [] and the backward compatible extension to 3GPP SCM [18]. Parameters used in the propagation are taken from Wireless World Initiative New Radio (WINNER) project and 3GPP TSG-RAN WG1. BS RN links are modeled as line of sight (LOS) and non-los (NLOS). Furthermore the antenna type used in the repeaters is normally assumed to be omni-directional antenna however it is interesting to consider directional antenna directed towards base stations and omni towards users. 3

Chapter LONG TERM EVOLUTION (LTE).1 3GPP LTE Overview The growing commercialization of Global System for Mobile Communications (GSM) and its evolution such as Universal Mobile Telecommunications System (UMTS) with High Speed Packet Access (HSPA) have been the focus topic of 3GPP. The GSM / UMTS system is perhaps the most successful communications technology family and its evolution to beyond 3G becomes important issue for the next global mobile-broadband solution. In parallel to evolving HSPA, 3GPP is also specifying a new radio access technology in Release 8 known as LTE in order to ensure the competitiveness of UMTS. LTE focuses to support the new Packet Switched (PS) capabilities provided by the LTE radio interfaces and targets more complex spectrum situations with fewer restrictions on backwards compatibility. Main targets and requirements for the design of LTE system have been captured in [] and can be summarized as follows. Data Rate: Peak downlink rates of 1 Mbps and Uplink rates up to 5 Mbps for MHz spectrum allocation, assuming receive antennas and 1 transmit antenna at the terminal Spectrum: operation in both paired (Frequency Division Duplex / FDD mode) and unpaired spectrum (Time Division Duplex / TDD mode). Enabling deployment in many different spectrum allocations with scalable bandwidth of 5, 1, 15, MHz, and better efficiency (downlink target is 3- times better than release 6 and uplink target is -3 times better than release 6) Throughput: Mean user throughput per MHz is 3- times (downlink) and -3 times (uplink) better than release 6. Cell-edge user throughput is also expected to be improved by a factor for uplink and downlink Latency: Significantly reduced control-plane and user-plane requirements, i.e. less than 5ms in the transmission of an IP packet (user-plane), allow fast transition times of less than 1ms from camped state to active state (controlplane) Costs: Reduced CAPEX and OPEX including backhaul for both operators and users, and effective migration from previous release shall be possible. 5

One of LTE requirement, as previously described, is to reduce the costs by simplifying the radio architecture. Therefore the number of nodes and interfaces in the network shall be reduced and it means that the 3GPP LTE Radio Access Network architecture need to group user plane functionalities into one network node called evolved Node B (enb) [3]. The resulting radio architecture is commonly known as System Architecture Evolution (SAE) and is depicted on Figure.1 below. Figure.1 E-UTRAN overall architecture [3] As shown in the figure, the 3GPP LTE Radio Access Network (RAN) architecture is different from the one of the previous 3GPP releases. The main difference is that a significant part of the radio control functionality has been distributed to the so-called enbs. Thus, it is possible to reduce latency with fewer hops in the media path and distribution of processing load into multiple enbs.. LTE Transmission Schemes 3GPP-LTE introduces the air interface access technologies from the use of orthogonal frequency division multiplexing (OFDM), multiple antenna technologies as well as modifications to the network architecture. OFDM is used in the downlink transmission and Single Carrier FDMA (Frequency Division Multiplex Access) technology is applied in the uplink transmission...1 OFDM OFDM is a multi-carrier transmission technique where the available spectrum is divided into multiple carriers, called sub-carriers. It is a modulation technique providing a high degree of robustness against frequency selectivity of transmission channels and achieving high data rate without inter symbol interference (ISI). The idea was proposed in mid 6s and used parallel data transmission and frequency division multiplexing (FDM). OFDM is a current well-established technology, for 6

example in standards such as Wireless Local Area Network (WLAN), Worldwide inter-operability for Microwave Access (WiMAX), High Performance Radio LAN (Hyperlan-), Digital Video Broadcasting (DVB) and Digital Audio Broadcasting (DAB) [9]. The utilization of OFDM in 3GPP LTE downlink transmission scheme enables additional benefits such as access to the frequency domain, i.e. enabling additional degree of freedom to the channel-dependent scheduler compared to HSPA, scheduling, power allocations, flexible bandwidth allocations, broadcast/multicast transmissions [8]. The idea of OFDM is to transform high data rate stream into low data rate streams transmitted in parallel in order to transform a frequency-selective fading channel into a set of frequency non-selective fading channels [9]. It uses relatively large number of narrowband subcarriers which are overlapped and orthogonal to each other. This is enabling OFDM to avoid the use of high speed equalization and to combat impulsive noise, and robustness against multipath fading as well as fully use the available bandwidth. Any type of non-ideal transmission channels spread the OFDM symbol causing the OFDM blocks to interfere one another. This type of interference where two adjacent blocks overlap causing symbol distortion is called Inter-Symbol Interference (ISI). One possible approach to combat this interference was to introduce a silence period between the transmitted frames, known as zero prefix. The silence period consists of a number of zeros added to the front of each symbol. The effect of ISI is still there however it is affected these prefix and they will be discarded in the receiver before demodulation of useful signal. Unfortunately, the zero prefix approach will destroy the periodicity of the carrier. Therefore instead of using silence period, one could extend the OFDM block by a cyclic prefix interval. The cyclic prefix interval consists of the last L samples of the OFDM symbol that are copied in the beginning of data block and it should be larger than maximum delay spread of the channel (i.e. T g T m ). If the cyclic prefix interval (T g ) spans more than maximum delay spread of the channel (T m ), the interference is entirely absorbed by the cyclic prefix which is then discarded in the receiver. The cyclic prefix facilitates the receiver s carrier synchronization and maintains the carriers periodicity because some signals are transmitted instead of a long silence period in the zero prefix approach... SC-FDMA The drawbacks of OFDM modulation are the large variations in the instantaneous transmitted signal power, high Peak to Average Power Ratio (PAPR), high sensitivity to frequency offset and a need for an adaptive scheme to overcome spectral nulls in the channels. The smaller PAPR of the transmitted signal, the higher average transmission power can be given for a given power amplifier. Therefore it is required to have expensive and inefficient power amplifier at the transmitter. This is very critical aspect in the uplink, since mobile terminal is power-limited and lowercost manufactured. In the other hand, SC-FDMA is basically another way to combat frequencyselective channel which delivers similar performance with essentially the same overall complexity compared to orthogonal frequency division multiple access 7

(OFDMA) system. 3GPP LTE proposes SC-FDMA transmission scheme in the uplink. SC-FDMA is a single carrier transmission based on DFT-spread OFDM where a block of N modulation symbols is applied to N-point DFT. The DFT spreads data symbols between all available subcarriers, combined with pilot symbols in time division multiplexing (TDM) and then mapped to proper subcarriers. After the N- point DFT, a size of M-point IDFT is applied to the signal, where M > N and the unused inputs of the IDFT equals to zero. At the receiver, the process is the opposite way in which after the M-point DFT is applied, the signal is frequency domain equalized and then the signal is finally converted into time domain using N-point IDFT. A comparison of SC-FDMA and OFDMA is drawn in Figure. which summarizes the difference in block diagram. {x n } N- point DFT Subcarrier Mapping M- point IDFT Add CP / PS DAC / RF Channel {x n } Detect N- point IDFT Subcarrier De-mapping/ Equalization M- point DFT Remove CP RF / ADC Figure. A block diagram of SC-FDMA ( + ) and OFDMA ( ) SC-FDMA has disadvantages in handling signal due to radio-channel frequency selectivity, but it can be solved in the enbs by utilizing more resources. One example of this method is to employ different forms of equalization at receiver however it requires higher receiver complexity. The other disadvantages of SC- FDMA are [11]: In SC-FDMA, the noise is averaged over all the bandwidth because the detection is done after equalized signal is reverted to time domain by IDFT, but in contrast, OFDMA performs the detection individually on each subcarrier Maximum Likelihood Detector is not feasible for Multiple-Input Multiple- Output (MIMO) Additional DFT processing increases mobile station complexity Unlike OFDMA, Localized SC-FDMA cannot exploit full advantage of multiuser diversity Distributed SC-FDMA has some issues, e.g. vulnerability to Doppler and frequency offset, and pilot design Only TDM pilots can be supported by SC-FDMA Low flexibility in multiplexing uplink control and data channels Degraded link-level performance as compared to OFDMA. The performance of OFDMA and SC-FDMA has been compared in terms of spectral efficiency (i.e. bits/s/hz) [7] in which two antenna configurations are used: Single-Input Single-Output (SISO) and 1x Single-Input Multiple-Output (SIMO). The simulation results show that SIMO makes the performance of SC-FDMA comparable to OFDMA but in the SISO case OFDMA outperforms SC-FDMA. 8

.3 Frame Structure In the time-domain structure, LTE transmission has frame length of T frame = 1 ms consisting of ten equally sized subframes of length T subframe = 1 ms [8]. One subframe consists of two equally sized slots of length T slot =.5 ms and each slot consists of a number of OFDM symbols including cyclic prefix. To provide consistent and exact timing definitions, different time intervals within the LTE radio access specification can be expressed as multiples of a basic time unit T 1 s = sec. Therefore T frame 37 and T subframe can be expressed as (37.T s ) and (37.T s ) respectively. The illustration of LTE time-domain frame structure is drawn in Figure.3. 1 frame; T frame = 1 ms...... 1 subframe; T subframe = 1 ms # #1 #9 1 slot; T slot =.5 ms Figure.3 LTE time-domain frame structure One slot of length.5 ms consists of six or seven OFDM symbols depending on cyclic prefix type. LTE defines two cyclic prefix lengths: normal cyclic prefix and extended cyclic prefix which corresponds to seven and six OFDM symbols per slot respectively shown in Figure.. One subframe = Two slots T slot =.5 ms Normal CP T C T U 66.7 µs = Extended CP T CP- T U 66.7 µs = Figure. LTE subframe and slot structure 9

The reasons of defining two cyclic-prefix lengths for LTE are described below [8]. a) Extended cyclic prefix may be beneficial in specific propagation scenarios with very extensive delay spread, for example in very large cells. In this scenario the additional robustness to radio channel dispersion is provided however it is less efficient from an overhead point-of-view. Therefore it is a tradeoff to choose which cyclic prefix length should be used. b) In case of Multi-Media Broadcast over a Single Frequency Network (MBSFN)- based transmission, the extended cyclic prefix is typically needed to cover both the main time dispersion part from the actual channel and the main timing difference part between the transmissions received from the cells. In the frequency-domain structure, LTE subcarrier spacing has been chosen to Δ f = 15kHz which corresponds to a useful symbol time T u 66.7 µs (8.T s ). A group of 1 consecutive subcarriers is called one resource block and it corresponds to 18 khz. The illustration of LTE frequency-domain structure is shown in Figure.5. One resource block (1 subcarriers) f = 15...... Figure.5 LTE frequency-domain structure The basic parameters of the LTE downlink and uplink transmission scheme are chosen to be aligned as much as possible. In the time-domain structure, the resource block for downlink and uplink is similar to illustration on Figure.3 and Figure.. The same basic parameter is also applied in frequency-domain structure as illustrated in Figure.5 however there is an unused DC-subcarrier in the center of the spectrum for the downlink case. The reason why it is not used for any transmission is that the transmission may coincide with the local-oscillator frequency at the BS transmitter and/or MS receiver, and may cause to unproportional high interference. In the other hand, uplink transmission has no unused DC-subcarrier because of single-carrier transmission and the presence of a DC-carrier in the center of the spectrum would have made impossible to allocate the entire system bandwidth to a single mobile terminal while still keeping the low-par single-carrier property transmission [8]. Thus the total number of subcarriers on a downlink carrier, including the DCsubcarrier, equals: N SC = 1.N RB + 1, where N RB is the number of resource blocks. And in the uplink transmission, the total number of subcarriers is N SC = 1.N RB.. Scheduling in LTE LTE transmission scheme uses shared-channel transmission in which the timefrequency resource is dynamically shared between users. This is similar to the 1

approach taken in HSDPA, although the realization of the shared resource differs between them, time and frequency in case of LTE and time and channelization codes in case of HSDPA. Figure.6 Channel dependent scheduling [8] The scheduler controls, for each time instant, to which users the shared resources should be assigned. It also determines the data rate to be used for each link that is rate adaptation. The scheduler is a key element and to a large extent determines the overall downlink performance, especially in a highly loaded network. Both downlink and uplink transmissions are subject to tight scheduling. It is well known that a substantial gain in the capacity can be achieved if the channel conditions are taken into account in the scheduling decision, so called channel-dependent scheduling. This is exploited already in HSPA where the downlink scheduler transmits to a user when its channel conditions are advantageous to maximize the data rate. Furthermore LTE has also access to the frequency domain in addition to the time domain. Therefore, the scheduler can select the user with the best channel conditions for each frequency region. In other words, scheduling in LTE can take channel variations into account not only in the time domain, but also in the frequency domain. This is illustrated in Figure.6. The possibility for channel dependent scheduling in the frequency domain is particularly useful for low terminal speeds when the channel is varying slowly in time. Channel dependent scheduling relies on channel quality variations between users to obtain a gain in system capacity. In LTE, the scheduling decisions can be taken as often as once every 1 ms and the granularity in the frequency domain is 18 khz. This allows relatively fast channel variations to be tracked by the scheduler...1 Downlink Scheduling In the downlink, each terminal reports an estimate of the instantaneous channel quality to the base station. These estimates are obtained by measuring on the reference signal, transmitted by base station and used also for demodulation purposes. Based on the channel-quality estimate, the downlink scheduler can assign resources to users, taking the channel qualities into account. In principle, a scheduled terminal can be assigned an arbitrary combination of 18 khz wide resource blocks in each 1 ms scheduling interval. 11

Figure.7 Downlink resource block assuming normal cyclic prefix 1.. Uplink Scheduling The LTE uplink is based on orthogonal separation of users and it is the task of the uplink scheduler to assign resources in both time and frequency domain (combined TDMA/FDMA) to different users. Scheduling decisions, taken once per 1 ms, control which mobile terminals are allowed to transmit within a cell during a given time interval, on what frequency resources the transmission is to take place, and what uplink data rate (transport format) to use. Note that only a contiguous frequency region can be assigned to the terminals in the uplink as a consequence of the use of single-carrier transmission on the LTE uplink. Channel conditions can be taken into account in the uplink scheduling process, similar to downlink scheduling. However, as will be discussed in more detail, obtaining information about the uplink channel conditions is a non-trivial task. Therefore different means to obtain uplink diversity are important as a complement in situations where uplink channel dependent scheduling is not used..5 LTE-Advanced Initial deployment of 3GPP-LTE is coming to the realization and the 3GPP is already planning to the future work for LTE, referred as LTE-Advanced [6]. LTE Advanced will be based on LTE, i.e. improvement of LTE in 3GPP Release 8, with a main driver of ITU requirements so called IMT Advanced G wireless system [1]. LTE Advanced targets higher data rates, reduced delay and latency, improved capacity and coverage, and fulfilling or surpassing the requirements for IMT Advanced. IMT Advanced is a concept used by ITU for radio-access technologies in mobile communication systems with capabilities beyond IMT-. The candidate technologies for IMT Advanced have been proposed to ITU and 3GPP already initiated the definition of requirements as well as technology components on LTE Advanced to meet all the requirements of IMT Advanced. 3GPP plans to submit to ITU on September 9 which will probably be fully specified in 3GPP release 1, and IEEE is likely to submit based on 8.16m standard which is an evolution of 1 Figure is adapted from: http://www.ericsson.com/technology/whitepapers/lte_overview.pdf 1

8.16e. The preliminary requirements for IMT Advanced can be found on the ITU s IMT Advanced website. It is quite certain that in terms of spectrum and bandwidth, all technologies beyond 3G such as LTE, LTE Advanced, WiMaX, etc., should have wider bandwidth and be backwards compatible. It means that there is no significant impact at user terminal while deploying LTE Advanced in the spectrum occupied by LTE now and the consequence is that the LTE Advanced should have a similar network as LTE network. The other compatibility is spectrum which is important for a smooth, low-cost transition within the network. Apart from the requirement on backwards compatibility, LTE-Advanced should fulfill the other requirements of IMT Advanced which are capacity, data rates and low-cost deployment. The final target for peak data rates leaves the possibility to have up to 1 Gbps in the downlink and 5 Mbps in the uplink, 1 Mbps for high mobility and 1 Gbps for low mobility. This is the headline requirements for G which is nailed to the same goal as in 3G: the growth in single-user peak data rates. However it is more important to provide high data rates over a larger part of the cell rather than for a single-user data rates. From the link performance perspective, the current cellular system in LTE is already close to the Shannon bound limit. And from a pure link budget point of view, the peak data rates targeted by LTE Advanced will require a higher Signal to Noise Ratio (SNR) than what is typically experienced in cellular networks. Although some link improvements are still possible, i.e. using additional bandwidth to improve the coding/modulation efficiency, it is also necessary to introduce concepts and tools to improve the SNR. A set of components and technologies being considered for LTE Advances includes: A. Wider-band transmission and spectrum sharing B. Multi-antenna solutions C. Coordinated multi-point transmission D. Repeaters. The improvement of SNR with denser deployment could be done by deploying denser number of base station or with different types of relaying solutions. The idea is to reduce the transmitter to receiver distance in order to get higher data rates and depending on the schemes, different types of relaying solutions can be employed. The detail of repeater is discussed in the following section. ITU-R IMT-Advanced website: www.itu.int/itu-r/go/rsg5-imt-advanced 13

1

Chapter 3 REPEATER CONCEPT 3.1 Repeater Overview The fundamental question occurred to wireless system area is how the signals should be distributed to and collected from the user terminal in very efficient manner to have high data rate and system coverage requirements. This developments yield architecture of the present cellular networks which can not meet the requirements of high data rate for the fourth generation (G) cellular systems. The extreme solution to solve this problem is to use larger bandwidth/spectrum or to significantly increase the density of base stations. Unfortunately, there is no indication that significant new spectrum will be available in the near future. In the other hand, high number of base stations results a considerably high deployment costs and it does not seem economically justifiable [1] [16]. The concept of relaying has been studied as a theoretical problem from a network information theory perspective in 197 s and in early 198 s. Capacity regions of simple relaying channels have also been evaluated [15]. However there was no further analytical study most probably due to the fact that there were no foreseeable applications at that time. Recently the concept has been brought up again to be used and proposed in the cellular network system. Relaying technique is a promising solution to substitute BS because a relay costs considerably lower than a BS but it acts to improve the link budget [13] just as a BS does. It has been used typically for handling areas with low signal coverage such as in radio shadowed areas, tunnels, business and industrial buildings, and where the traffic is too low to justify the installation of a base station. There are many products available in the market for this type of solution, e.g. [7] [8]. Several types of relay which have been developed in cellular network system are: Amplify-and-forward (AF) relay or sometimes referred as repeater (Layer 1 relay) Decode-and-forward (DF) relay or Layer relay Self-backhauling relay or Layer 3 relay. A repeater (L1 relay) is a relay which amplifies the received signal and forwards it to user. The repeaters are transparent to system and they are blind when forwarding the signal without knowing whether it is desired signal, interference or noise. 15

According to what frequency the signals are transmitted, the repeater is classified into two types: on-frequency repeater and frequency shifting repeater. The explanation of these repeaters is given in section 3.1.3. DF relays, in the other hand, decode the received signal first before forwarding it to user. Hence there is additional delay occurred due to signal processing and it needs other resources for the relays to MS links, i.e. first time/frequency slot for BS to MS link and second time/frequency slot for relay to MS link. The advantage of this type of relay is of course no noise forwarded by relay, however it is difficult to receive both the composite and desired signal at the same resource block in LTE therefore it needs more resources. Another type of relay is L3 relay which is often denoted as self-backhauling and performed on layer 3. Self-backhauling is recently attracted researchers because it is very similar with Layer relaying in their basic characteristics and it is proposed to be one of the feature technology in LTE-Advanced [6]. This thesis only studies on-frequency repeater which amplifies the received signal before retransmission. 3.1.1 Basic Repeater Design Repeater basically consists of two way amplifier with duplex filter as shown in Figure 3.1. It has two antennas: one for the connection to the parent base station and the other one for service area to the users. They are called donor antenna and service area antenna respectively. The repeater receives signal from donor antenna, filters the signal, amplifies the signal and directs to the other antenna to be transmitted. The repeater is located in the cells and is used to improve the cell coverage and cell capacity in certain areas. Physically a repeater could be in one package with two built-in antennas inside or it could be a box with two built-out antennas connected with cable/fiber. The second implementation is sometimes preferable to increase the antenna isolation and to reduce feedback interference by giving more distance to both antennas. Donor Antenna Service Antenna Adjustable amplifiers Filters Filters Adjustable amplifiers Figure 3.1 Repeater Block Diagram 3.1. Antenna Isolation Antenna isolation is an essential issue for the performance of a repeater because the feedback signal from service area antenna to donor antenna acts as interference. An illustration for antenna isolation is drawn in Figure 3.. In the indoor installation the isolation between the donor antenna, generally mounted outside the building, and the 16

service area antennas inside the building is not a big issue. The attenuation path is high enough due to concrete, walls and the usage of feeder cable, but in the outdoor installation the attenuation path from service to donor antenna is relatively low. If the antenna isolation does not meet the requirement, the repeater acts as an oscillator making the repeater itself not to work. RBS Coverage Antenna Antenna Isolation = Repeater Gain + 15 db Margin Donor Antenna Service Antenna Repeater Figure 3. Antenna Isolation Self-oscillation can be avoided if the overall path loss between the repeater service antenna and donor antenna, referred as antenna isolation, is higher than the maximum repeater gain (G R,max ). An additional 15 db margin is considered to ensure the antenna isolation according to specification for UTRA Repeater []. As a rule of thumb, the antenna isolation should fulfill the following inequality: Antenna isolation G R,max + 15 db (3.1) In the implementation antenna isolation must be resolved before a repeater is installed. If antenna isolation cannot be reached, it is necessary to decrease the maximum repeater gain at the expense of losing repeater coverage area. There are several factors that influence the antenna isolation [17]: The distance between antennas: the path loss is higher when the distance between antennas is increased. It is roughly proportional to the square of the distance between the antennas (free space propagation loss) and in the UMTS frequency system, antenna separation up to 1 m is recommended Antenna lobe width and front to back ratio: higher antenna isolation can be reached if the antennas have narrow lobe width and higher front to back ratio. The service area antenna is typically an omni directional antenna, but a directional antenna can be employed especially at donor antenna to increase antenna isolation. The typical donor antenna gain ranges from 15 to 18 dbi Antenna polarization Shielding Surrounding environment: the reflection and attenuation properties of all materials close to antennas can influence the antenna isolation drastically Isolation through Self Interference Cancellation 3. 3 Peter Larsson, MIMO On-Frequency Repeater with Self-Interference Cancellation and Mitigation, VTC9-Spring, 9 April 9. 17

3.1.3 On-Frequency and Frequency Shifting Repeaters An on-frequency repeater, as the name clearly explains, is a repeater which uses the same frequency band on both base station to repeater link and repeater to user terminal link. On-frequency repeater which receives and transmits on the same frequency band sometimes creates problems since it relies on the antenna isolation which effectively limits the repeater gain. Therefore it is important to choose high gain antennas with low side/back lobes and performing a good antenna installation. The resulting coverage from an on-frequency repeater strictly depends on the antenna type and installation skill. Generally it is difficult to obtain high antenna isolation between donor and service area antenna at the repeaters. Indeed, as shown in Figure 3.3, on-frequency repeater needs to have an additional and costly component to achieve sufficient antenna isolation known as canceller, therefore it makes the entire repeater solution approach more expensive and less desirable. Instead, frequency shifting repeaters can be used. These repeaters use one frequency to communicate with base station and use another frequency to talk with mobile station. For example, it converts frequency F1 to frequency F for downlink, and vice versa for the uplink communication. Therefore, sometimes frequency shifting repeaters are also known as frequency translation or conversion repeater. Feedback Interference F1 Antenna isolation (3dB) F1 Canceller (5dB) Base Station On-Frequency Repeater Mobile Station F1 (a) Feedback Interference F1 F o Antenna isolation (3dB) Filter (5dB) Base Station Frequency Shifting Repeater Mobile Station (b) Figure 3.3 On-frequency (a) and Frequency Shifting (b) Repeaters F F1 Feedback interference Frequency shifting repeater uses different frequency for two hops channel which can solve the antenna isolation problem. It is possible to provide up to 1 db repeater gain with much lower antenna isolation requirement. In general, frequency shift repeaters shall be more cost-effective solution than on-frequency repeaters because they do not have canceller and only have band pass filter in the RF module. The high gain and low antenna isolation requirement of this unit are suitable for coverage extension in rural areas. However frequency shift between link between repeater to user and link between base stations to repeaters requires more resources in order to allow the repeater to work correctly and this would result a reduced capacity (duplex loss). Obviously more complex frequency planning and guard channels are needed. Therefore frequency-shifting repeaters are not recommended to be used widely in LTE network 18

and successful installation enforces high requirements on low adjacent channel signal levels in the area. The other common problem that occurs in the frequency-shifting repeater is synchronization problem. Even though a repeater is equipped with a very high quality local oscillator, a frequency drift in time may not be avoided. A manual frequency re-tuning with an interval of period time is necessary, i.e. less than two years for repeater in GSM. 3.1. Repeater Delay In two-hop communication model, the repeater introduces delay due to filtering, processing and feeder links used in application which could deteriorate the signal coming to user terminal. As long as the delay is within cyclic prefix, the signal can be recognized and combined to obtain the original transmitted signal. This happens because OFDM scheme employs cyclic prefix, a circular extension of the data symbol to combat inter-symbol interference introduced by the frequency selectivity of the radio channels. The cyclic prefix is used in order to avoid the effect of time dispersion by multipath propagation which makes two consecutive frames to interfere each other at the receiver. If the frame delay is more than cyclic prefix, the receiver can not combine the signal and they are seen as interference. Therefore it is important to keep the repeater delay as low as possible such that the total delay at user terminal must be within cyclic prefix period. An example of total delay in frequency selective repeater in GSM network is - 6 μs while the extended cylic prefix period in LTE is 1.857 μs. Another effect of repeater delay is causing the positioning system to estimate the UE position wrongly. This effect depends on what positioning system type is used. In Round Trip Time (RTT) and Observed Time Difference of Arrival (OTDOA) methods the estimation of UE position strongly depends on propagation delay measurements between the UE and NodeB, however in Cell-ID method the repeater delay will not affect the identification of the serving cell. 3.1.5 Interference and Capacity The deployment of repeaters in cellular network has a main task to extend the cell coverage while keeping the total investment costs lower than the deployment of NodeBs without repeaters. Repeaters have some impacts on the cellular system in term of interference, capacity and coverage especially in outdoor scenarios. In rural areas where the traffic density is low, the number of NodeBs is generally dimensioned according to coverage. In such applications where the capacity is not limited, the repeaters are used to replace one or more NodeBs as long as the service quality is still acceptable. In urban areas capacity is an important factor for determining the number of NodeBs and it is limited by the downlink power. A repeater in this area generally increases the overall downlink power because it amplifies both the desired signal and interference. Thus it actually increases the overall downlink capacity if the desired signal without any interference is amplified. However, in multi-cell environment inter-cell interference exists and it means that there is a tradeoff between the interference and capacity in the system. 19

3.1.6 Repeater Applications Choosing a base station or a repeater to be deployed is not an easy way but it is possible to provide general rules in order to make one choice more sensible than the other one. A base station basically needs high electrical power feeding, a physical transmission link to the core network, and it is relatively expensive. On the other hand a repeater is a cost-effective complementary solution for solving coverage problems, but it does not add any new hardware capacity to the system and it has often somewhat worse radio performance than a base station. Implementation of a repeater solution starting from the coverage extension issue to a complete working system is a very rapid process. Acceptable signal strength at the donor antenna and a capacity of the donor cell which allows serving the new MSs introduced by the repeater are some prerequisites required for a repeater solution. The repeater can be seen as a natural way in expanding the network coverage, therefore there is an expectation of increased capacity demand in the system. In summary, the cost, required capacity, and limited implementation time are critical factors to this process. Repeater Service area antenna Donor antenna User Terminal Base Station (a) (b) Figure 3. A typical repeater s installation in outdoor scenario (a) and indoor scenario (b) Typical repeater applications are to improve cellular system performance in outdoor or indoor scenario. The performance could be system coverage and/or capacity. Figure 3. shows a typical outdoor repeater on the building s rooftop and indoor repeater. The outdoor scenario is typically used for macrocell coverage extension purposes where the direct signal path from BS to MS is obstructed by hills or there are coverage holes in the service areas. Note that in this case the capacity

should not be a limiting factor. The indoor scenario is intensely used for in-building coverage where the external macrocell coverage is not sufficient. 3. Two-Hop Communication Model The capability of relaying method is perhaps a promising architecture to be implemented in the wireless network structure as a cost efficient solution. It allows coverage extension with high data rates and can reduce the total number of expensive base stations in the system. The multi-hop feature also allows many units to retransmit the signal to the destination in several hops, but two-hop relaying could decrease the operation cost and simplifies the routing function. Furthermore fixed relaying is utilized to simplify the radio protocols and to improve the efficiency. Two-hop fixed relaying method is based on fixed repeater node (RN) deployed in the system infrastructure and the number of hops allowed is restricted to two, i.e. BS to RN and RN to MS. This simplification further increases the practical implementation of the multi-hop technology and reduces the complexities. Figure 3.5 shows the layout of our model: hexagonal cells with radius R, one cell site with three sectorized antennas BSs, and repeater node (RN) placed at some distance from the base station to the cell border. Furthermore, 3 repeaters per cell are considered in our work. : Base Station : Repeater Node : Mobile Station Figure 3.5 Cellular layout of the system A simple illustration of two-hop communication model and how the system works for a single cell is drawn in Figure 3.6. The desired signal is sent from serving base station to the user (black line) and the interference signal comes from other base stations to the user (red line). The blue line represents the desired signal sent through repeaters in other cells. The received signals (with thermal noise) at repeater is amplified and retransmitted to the user, and then all signals received at user terminal are summed together as multipath signals. 1

1 Mobile Base Station Base Station Repeater Figure 3.6 Simple illustration of two-hop communication model Desired signal Desired signal from outer repeater Interference from outer cell If C, is the path gain from base station-k of cell-i, E s is the symbol energy, and i k σ z is the noise variance, the SINR equation of a user in cell-k in this model is: Γ = z N repeaters + C K, k k = 1 N N repeaters cell i= 1 k = 1 i K where thermal noise is assumed to be white. σ C E s i, k E s (3.) 3.3 Advanced Repeater As being explained in Section.5, layer 1 relays or also referred to repeaters are considered to be one of potential technology features of LTE Advanced. Some problems appeared to repeaters in 3GPP Rel-8 are energy consumption, interference problems, lower throughput when served by repeaters, difficult to monitor operation and potential features for improved performance. Practically the conventional repeaters are continuously even when there is no data transmitted, so the repeater consumes more energy than what it needs. Moreover in multicells scenario the repeaters may contribute more interference to the other users and could degrade the overall performance. Therefore it is important to have some advanced functionalities, i.e. to switch the repeater on/off in an optimum way, to have frequency selectivity and gain controllability, to use multi antenna ability and advanced antenna processing, etc. There are two advanced repeater functionalities analyzed in this thesis. First, an advanced repeater may use time and/or frequency selective functionality and it only

forwards when there is data to be forwarded and chooses the frequency bands needed. Second, it may control the repeater gain. In these ways, power consumption and interference can be reduced. Furthermore power control at repeaters can give benefits in order to reduce unnecessary power and to increase the overall performance. The simplest power control is to have an on/off switch that will turn the repeater on if there is data to be forwarded from serving base station. More complex power control algorithm can control repeater gain such that all users in the system get the best performance with fairness condition and satisfaction. This power control and other advanced repeater functionalities are not in the scope of this thesis hence they may be considered as future work. 3.3.1 Frequency Selective Repetition Frequency selective repetition is one of the advanced repeater functionalities which can improve the system performance. The frequency selective functionality is used in order to reduce the unwanted signal being forwarded to the users therefore the interference introduced to other cells is reduced. This method is similar to Inter-Cell Interference Cancellation (ICIC) where MSs in the the cell edge served by repeater on coordinated resources and BSs control the repetition of frequencies used by repeater. Hence it can allow an improvement of the user s SINR and system performance in general. Figure 3.7 illustrates this functionality in a simple case. Figure 3.7 An illustration of coordinated frequency selective repetition in Uplink and Downlink The frequency selective repetition is basically implemented in the frequency selective repeaters by employing filters. According to the bandwidth of the filters implemented, these repeaters can be classified into: broadband repeater, band selective repeater and frequency selective repeater. An illustration of these repeaters with different type of filters is given in Figure 3.8. Broadband Band Selective Frequency Selective f Figure 3.8 Repeaters with different filters f f 3

Frequency selective repeater has a bank of band-pass filters to amplify multiple channels or resource blocks as shown in Figure 3.9. The bandwidth of a repeater can be made as minimum as possible to achieve the maximum selectivity against adjacent channels, i.e. the minimum schedulable bandwidth in LTE system is one resource block. However sharper filter will cost more and makes the processing time longer. It is sometimes assumed that the filters have the same bandwidth and component. Therefore if the bandwidth of a single filter is narrower, the number of filters in a repeater is higher and the granularity of the filter bank is also finer. Controller Figure 3.9 Controllable filter banks in the repeater In LTE the frequency selective repeater can be used to retransmit the data to users scheduled in the resource blocks. Each user can be scheduled in a consecutive resource blocks or in a localized resource blocks. An illustration of repeater frequency selectivity function and the amplification in frequency domain is drawn in Figure 3.1. Note that the repeater is also potentially controllable in time domain, i.e. it may be active when there is resource block assigned to active user served in the cell and can be off when there is no resource block used by active users. Scheduled resource blocks User 1 User Frequency Conventional repeater Frequency Frequency selective repeater Frequency Figure 3.1 Illustration of frequency selective repeater 3.3. Repeater Gain Controllability Repeater gain controllability is another feature in the advanced repeater which can be used in order to reduce interference, to reduce power consumption, to improve uplink

power control and to avoid instability. This feature is motivated from a simple illustration where a repeater could be on/off or continuously transmitting in time. On/off state simply refers to if there are data to be forwarded or not. Furthermore different amplification factor or gain may give different system performance because higher gain contributes higher interference to the system. There are two granularities that can be controlled which are time granularity and frequency granularity. In the time granularity, the repeater may use long time scale or at scheduling rate to review and to change the gain based on power control algorithm. We define fast repeater as a repeater allowed to change the gain drastically at the next time instant and slow repeater as a repeater which needs to increase/decrease stepwise. In frequency granularity case, a repeater may amplify the entire bandwidth or in a frequency-selective way. We define fine resolution as filters correspond to one resource block and coarse resolution as filters correspond to a chunk of resource blocks (e.g. 5 MHz). The tradeoff of this functionality is the need of synchronization and control signaling from base station to repeaters and it makes the interference estimation accuracy decrease. In our simulation, there are five repeater states which will be considered. They are:, slow coarse resolution, slow fine resolution, fast coarse resolution and fast fine resolution. The example of gain control functionality is given in Figure 3.11. Figure 3.11 Example of gain control functionality 5