Introduction Long Term Evolution-Advanced (LTE-Advanced) is the 3rd Generation Partnership Project (3GPP) candidate technology which is expected

to enhance cell edge capacity, system throughput as well as reduce the user and control plane latencies. It offers 1Gbps in the downlink (DL) and 500 Mbps in the uplink (UL). Moreover, it will support the extended carrier bandwidths upto 100 MHz. In order to achieve this performance level, LTE-Advanced component technologies have been studied namely as Carrier aggregation (CA), Extended multiple-input multiple-output (MIMO), Relay node (RN) deployment, Femtocells deployments, Coordinated multipoint transmission and reception (CoMP). In this course, the main task is to study the effect of relay nodes deployment within the existing heterogeneous network. Similarly, it also required to discover the losses and gains for Relay enhanced cell (REC) model such as the deterioration in the SINR levels and improvements in network throughput levels. While deploying the relay nodes, some issues also arise related to RN positions, resource scheduling and user equipment (UE) performance.

Paper 1: Effect of Relaying on Coverage in 3GPP LTE-Advanced

Research Problem
Generally in the wireless networks, the cell edge users are experiencing a low signal-to-noiseinterference ratio (SINR), causes the low user througput which leads to overall bad system performance. Similarly, it also causes the small cell coverage and capacity at the cell edge. Moreover, the 3GPP LTE-Advanced is required to provide peak data rates in order to support the high data services and applications. In order to solve this problem, the relaying technique has been proposed. It is two hop techniques which are usually deployed within the macro base station coverage area, at the coverage holes and cell edge. Relays are expected to improve the system capacity and coverage as the low SINR users will handover to the relay node and utilize the system resources efficiently.

Research Methodology
Here the research methodology has been proposed to check the relay impact on system performance from the coverage perspective. In this case, a specific throughput level (10%-tile of throughput CDF) assumed as threshold, while relaying impact will increase the inter-site distance (ISD), which leads to the cell coverage improvements. The relay node transmission power and Iso-performance curve (also known as indifference map) are the performance metrics. The Iso-performance curve is the set of points, in which each point provides exchange ratio between the eNBs and RNs, promising the same system performance i.e. 10%-tile of throughput CDF. In order to calculate the iso-performance deployments, an iterative algorithm has been used. The output of this algorithm is the combination of RN with ISD combinations, while the reference scenario is ISD 500 meter with eNB only network. In each iteration, one RN is added per sector as well as ISD increased

3 upto level, that the new deployment admits the 10%-tile CDF throughput is equal to the reference deployment scenario. This process continues till the maximum numbers of RN per sector are determined. Similarly, Iso-performance (also known as indifference map) deployments are used to obtain a tradeoff between the number of RNs and eNBs. This trade-off is known as exchange ratio between eNB and RNs while keeping the same system performance level (i.e. 10%-tile of throughput CDF). Exchange ratio = _RN/ _eNB Here the numerator gives the required density of RNs while denominator is the reduction in density of eNBs due to relay deployment.

Main Conclusion
The investigation concludes that the UEs connected to the RN via access link, cannot achieve the maximum throughput, due to resources utilization and bottleneck by relay link. Similarly, for the case of 5 and 10 RN per sector with 24 dBm transmission power, have achieved better throughput level as compared to reference scenario. In addition to it, the 24dBm and 38 dBm RN transmission powers were compared, in which the latter case requires less number of RNs (4.6 eNB for 39 RN) as compared to former (4.6 eNB for 133 RN), to achieve desired system performance. The main factor for later case selection was saving high cost savings.

Paper 2: Business Impact of Relay Deployment for Coverage Extension in 3GPP LTE-Advanced
Research Problem
This paper discusses the business impact of relay node (RN) deployment on LTE-Advanced networks. Relay nodes are believed to give high data rates coverage with minimum operator cost. It also enhances the network capacity by increasing the overall cell throughput, due to efficient utilization of network resources.

Research Methodology
This work assumes an inband two-hop RN in coverage limited scenario (sub-urban). Moreover, the purpose of RN deployment is only, to enhance the network coverage and capacity rather then providing new services. The performance metrics used for studying the business impact are, Total Cost of Ownership (TCO) of different network scenarios (without RN and with RN) and Iso-performance scenarios. TCO model consist of several parameters, one time services like CAPEX (equipment cost), IMPEX (implementation cost, planning, acquisition and construction) and recurring services like OPEX (operation and maintenance cost). Similarly, Iso-performance (also known as indifference map) is used to obtain the exchange ratio between eNB and RNs while keeping constant the same system performance level. Exchange ratio = _RN/ _eNB Here the numerator gives the required density of RNs while denominator is the reduction in density of eNBs due to relay deployment. Three different price levels are assumed for site costs namely high, mid and low site costs. The eNB used Microwave (MW) or Fixed Network (FN) links to connect to network. Different types of RNs with different transmit power (i.e. 24, 33 and 38 dBm) and different architecture (Carrier grade and Consumer grade) are assumed. Carrier grade represents a system that is extremely reliable with proven capabilities and high availability. While the consumer grade are the similar to carrier grade, having low price and low availability.

Main Conclusions
It has been observed that the exchange ratios decreases for RNs with high transmission powers, mean that less number of high power transmission power relays will be needed to replace eNB in order to get same performance level. For the case, with one time cost (CAPEX and IMPEX), FN backhaul and low cost site, the one-time costs for all types of RNs are always lower than eNBs. The reason for this cost benefit is the high site costs of eNB which mainly include the civil work costs. While the RNs can be easily installed on the street lamp posts which required less site acquisition efforts. Moreover, the consumer grade architecture with low power RNs outperforms the carrier grade architecture due to lower equipment costs. In addition to it, operator can do 30% savings within period of 5 years by deploying RNs instead of eNB. Similarly, with OPEX, MW backhaul and high cost site, high transmission powers RNs are more favorable as compared to low transmission power. The main dominating parameter is OPEX as the operator will need high number of low transmission RNs as well as the O&M costs will be more than the eNB case. Nevertheless, operator can save more than 30% cost with high transmission power RNs.

Paper 3: On the Coverage Extension and Capacity Enhancement of Inband Relay Deployments in LTE-Advanced Networks
Research Problem
To fulfill the ITU requirements, 3GPP LTE-Advanced is required to provide high peak data rates in order to support the advanced services and applications. To meet these requirements, different LTEAdvanced technologies have been studied in which includes relaly node (RN) deployments. According to resource utilization on backhaul link (eNB-RN), relay nodes have been differentiated into different types. Inband RNs utilizes the same frequency spectrum for both the, backhaul link (eNB-RN) and access link (RN-UE). Both of these links time-divisioned multiplexed as both are operating on single frequency. This approach may create some limitations on the resource utilization at backhaul link of inband RNs which can be reduced by introducing enough physical isolation between the antennas structure of two links. Examples are Type 1 and Type 1b, while the difference them that, Type 1b added a physical isolation between the backhaul and access link antennas. In contrast to it, the outband RNs are using different frequency spectrum for both links. This type of relaying bring flexibility in resource utilization at backhaul link at the cost of deployment cost as separate extra frequency spectrum will be needed for backhaul link. Example is Type 1a. In this paper, the Type 1 and Type 1b RNs performance are investigated for different LTE-Advanced propagation scenarios in terms of network coverage and capacity. The purpose of this study is to examine the relaying behavior in different LTE-Advanced network environment as follow; RN deployment cost and gains Effect of inband backhaul link overhead. Comparing the gain of heterogeneous deployments (eNB+RN) with homogeneous (eNB only) deployments.

Research Methodology
In this paper, the Type 1 and Type 1b RNs deployments performance are compared in terms of network coverage and capacity. The performance metric used for network coverage is the exchange ratio between eNBs and RNs, gives the amount of RNs for one eNB, to achieve same network performance level. For coverage comparison, the paper 1 methodology has been used. Please refer to Paper 1. Similarly, average cell throughput levels give the network capacity gains. Below methodology is used for network capacity oriented comparison.

7 In two hop relaying, the UE end-2-end throughput ( ) can be calculated by;

In Type 1 relaying, the UE end-2-end throughput ( as;

) is bottleneck by the backhaul link, calculated

In Type 1b relaying, the backhaul and access link antennas are enough physically isolated. Hence, simultaneous transmission is possible on the both links as well as reduces the limitation on the backhaul link resources. The UE end-2-end throughput ( ) calculation can be done as;

Propagation Models: The radio environment has a large impact on wireless communication system performance. In the study, three different propagation model scenarios have been considered. In propagation scenario 1 (Sc1), all the links (Direct link (eNB-UE), Backhaul link (eNB-UE), Access link (RN-UE)) are using singles-slope channel models (e.g. Okumura-Hata models). It means that the donor eNB, RNs and UEs are at NLOS to each other. However, practically it is not possible as there is probability of line-of-sight (LOS) conditions for small cell sizes. In propagation scenario 2 (Sc2), a dual probabilistic channel models have been proposed where mixed LOS/NLOS modeling for Access link (RN-UE) will be used. The probability function will calculate the LOS/NLOS distance between the UE and RN and will use an appropriate channel model (LOS/NLOS) for the access link. While the Direct link and Backhaul link using the same single slope models. In propagation scenario 3 (Sc3), a dual probabilistic channel models are also used for Direct link and Backhaul link. Similarly, the channel models used for access link are the same as used in scenario 2 (Sc2). The network consists of regular hexagonal cellular layout with 19 tri-sectored macrocell sites. The simulation has been done for both 3GPP Case 1 (urban) and 3GPP Case 3 (sub-urban). For this study, 7 RNs and 14 RNs, which constitute, respectively, 1 and 2 tiers are deployed in the network. The first tier RNs are deployed at the cell edge in such a way that there should be minimum coverage gap between neighboring RN cells. Similarly, the second tier RNs are deployed near to the eNB. For indoor UEs, 20 dB penetration losses added on the access link and direct link. The UE link throughput is calculated by given formula as;

Main Conclusion
Coverage extension: ISO-Performance (ISD 500m). The results show that Type 1 and Type 1b have small ISO-performance difference for all three scenarios. Similarly, the required number of RNs in ISO-combination of scenario 3 is less than those in scenario 1 and scenario 2. When comparing scenario 1 & 2, both are using the same channel models for direct link and backhaul link, but the scenario 2 have good performance then scenario 1. The reason is LOS probability used on the access link give a good performance. Moreover, in case of scenario 2 & 3, the difference occurred due to good performance of direct link as the both the scenario have the same channel models for backhaul and relay link. Hence simulation results show that for scenario 3, both types of RNs gives cost-efficient solution for the coverage extension as compared to other two scenarios. ISO-Performance (ISD 1732m). It can be stated from the results that both types of RNs have the same kind of behavior from coverage perspective. While the system performance for all three scenarios are almost the same, as that for ISD 500m. Cell Throughput: ISO-Performance (ISD 500m). It is evident from results, that Type 1b RNs have better performance especially at high SINR levels as compared to Type 1 RNs in both RNs deployments types (i.e. 7 RNs and 14 RNs) for scenario 1. The reason is the backhaul link, which bottlenecks the e2e throughput of two-hop link. Moreover, the Type 1b RNs have good performance for scenario 2 as compared to other two scenarios while achieving average cell throughput gains. ISO-Performance (ISD 1732m). In this case, the propagation model of scenario 3 has been assumed. This is observed that Type 1 and Type 1b RNs have the same performance for both tier of RNs deployment (i.e. 5 RNs and 10 RNs) at the low percentile throughput CDF levels. Moreover, RNs deployments generate good results for scenario 2 and 3 as compared to scenario 1. The reason is that, all the UEs on the access link are assumed in NLOS conditions which brought severe propagation loss.

Paper 4: Performance Enhancement in LTE-Advanced Relay Networks via Site Planning

Research Problem
In 3GPP LTE-Advanced, relay node (RN) deployment enhances the network coverage extension as well as overall network throughput. It admits a wireless backhaul link (eNB-RN) towards the serving eNB, which provides significant deployment flexibilities. Rather than random deployment, an appropriate site planning is required; in order to achieve a good backhaul link performance by reducing the shadowing fading effects. The better, the SINR level of backhaul link, the better will be the RNs performance in the network.

Research Methodology
To find the optimal locations for RN deployment, two approaches has been proposed for it. First, the RN is allowed to connect with eNB with signal level. Secondly, the RN should be located in such place where it experiences a high shadowing from the interferer eNB or RN. To explain the above scenarios, two strategies have been proposed. A) RN cell selection (A1): No cell selection (A2): Cell selection allowed B) RN site location selection (B1): Only one RN location (B2): Multiple RN location From the above strategies, four approaches have been deduced for RN cell selection and site locations as follows; 1. 2. 3. 4. (A1, B1): No cell selection and No RN site location selection. (Reference case) (A2, B1): Cell selection allowed but No RN site location selection. (A1, B2): No cell selection but RN site location is allowed. (A2, B2): Both cell selection and RN site location allowed.

The performance metrics for the above approaches are SNR (signal-to-noise ratio) or SINR (signal-tointerference-noise ratio). Moreover, three different RN deployments have been assumed i.e. RN deployment at the cell edge region, the intermediate region and the cell center region. Access link is not taken into account as this work is to check the backhaul link performance.

10 The results have been generated for both 3GPP case 1 (ISD = 500m) and case 3 (ISD=1732m) with single RN in macrocell.

Main Conclusion
The system performance gains from all the above approaches are benchmarked with the reference case (A1, B1) which gives the achieved system performance. For 3GPP Case 1 (ISD 500 m), the results show that (A2, B2) gives a good performance especially at the cell edge. Here the RN is allowed to connect a best eNB as well as can be deployed at optimal locations as it will counter act the shadowing and interference effects from the interferer eNB. Moreover, it is also noticed that system performance is gradually decrease as the RN is moving from cell edge towards cell center. For 3GPP Case 3 (ISD 1732 m), the performance is almost the same of that of Case 1 with difference that system performance is faster decreases as the RN is moving from the cell edge towards cell center. Hence proper site planning can improve the system performance, by reducing the shadowing standard deviation by 2dB for 3GPP Case 1 and 2.9 dB 3GPP Case 3.

11

Paper 5: Performance of Amplify-and-Forward and Decode-andForward Relays in LTE-Advanced

Research Problem
The current broadband wireless networks are considering the large cell sizes. Due to which the cell edge users are experiencing low Signal-to-Interference-Noise-Ratio levels, which leads to lower cell edge throughput. In order to solve problems, the idea of relay node (RN) deployment at the cell edge was proposed. Relay node acting as node between the serving eNB and user equipment (UE) which receive a signal from the eNB and retransmit to the UE. In this paper, two different types of RN has been considered namely as Amplify-and-Forward (AF) and Decode-and-Forward (DF). These two RN approaches are needed to be compared due to its consideration for 3GPP LTE-Advanced study.

Research Methodology
Comparison of AF and DF has been done on the basis of UE throughput calculation, connected to relay nodes. 3GPP Case 1 (ISD 500m) with 2GHz carrier frequency in downlink has been assumed. AF is a full duplex relay node amplifies a signal received from the first hop and retransmit to the second hop. In AF RN, the transmit signal on the access link leaks to backhaul link receive antenna causing interference. This type of interference is known as Loop Interference (LI), due to which separate antennas with proper physical isolation are used for transmission and reception purposes. Moreover, AF relays do also amplify the interference and noise with desired signal which deteriorates the overall SINR level. In AF relaying, the SINR at UEs connected to RN consists of desired signal and interference and noise signals. Desired signal is the summation of useful signal from eNB plus signal from RN. Similarly, the interference and noise signal contains the LI, relayed noise and UE receiver noise. The UE SINR is used to calculate the system spectral efficiency (SE) by Shannons approximation formula.

Where and denote the signal to noise ratios of backhaul link and access link. Similarly, the refers to the signal to noise ratios of direct link. The physical isolation between relay antennas are given by . Similarly, DF is half duplex relay nodes decodes a received signal from first hope and then re-encodes and retransmit the pure decoded signal to second hop. It adds delay to the system due to encoding and re-encoding process of signals as well as increases the system complexity. In DF relaying, the throughput over backhaul and access links will be maximized, if both the links have the equal throughputs.

12 In addition to it, the access link resources can be reused by multiple RNs within the same macrocell. Hence the end-to-end spectral efficiency increases with n concurrent RN transmission but is bottleneck by the efficiency of the backhaul link.

Where is backhaul link spectral efficiency while efficiency.

refers to the access link spectral

Main Conclusions
The results show that the AF relaying efficiency has been degraded due to loop interference. The reason is lack of physical isolation between transmit and receive antennas at RN. Similarly, the DF RN shows good performance on the cell edge while AF RNs has been proposed to deploy in middle of the cell for high spectral efficiency. The AF RN performance is better then DF RN, if the DF RN have single transmission. But for the concurrent transmissions, the scenario is opposite i.e. DF RNs outperforms the AF RNs both in the middle as well as at the cell edge. Moreover, for different backhaul link gains, DF RNs accomplished better results as compared to AF RNs.