Research On Key Techniques Of Radio Resource Allocation In Relay-enhanced Cellular Networks

Increasing capacity along with coverage in conventional wireless cellular networkdictates dense deployment of base stations (BSs), which leads to substantial increase incapital and operational expenditure. To alleviate this problem, the relay-enhancedcellular network is proposed, which has attracted considerable research attention in bothacademia and industry, and become a new research focus. It’s noteworthy that, as one ofthe key technologies in mobile communication systems, radio resource allocation playsa very important role in fully playing the advantages of the relay-enhanced cellularnetwork. However, the introduction of relay stations has changed the networkarchitecture, so that the network environment becomes more complex and thus puttingforward new challenges for radio resource allocation. Therefore, the research on radioresource allocation in relay-enhanced cellular network has become an urgent and criticalissue. In this dissertation, under the guidance of information theory and convexoptimization, by the method of combining theoretical analysis and experimentalresearch, the cross-layer resource allocation techniques in relay-enhanced cellularnetwork are intensively studied.The rapidly growing mobile communication services impose higher requirementson both resource utilization efficiency and user fairness. To balance the efficiency andfairness of resource allocation, the utility-based resource allocation in multiuserOFDMA cellular networks with both soft QoS and BE users is studied. Firstly, autility-based cross-layer resource allocation model is proposed, with the objective ofmaximizing the system utility. Then, a dynamic subcarrier allocation algorithm, referredto as Mix-Max-Utility (MMU) is devised. Numerical results indicate that the MMUalgorithm not only guarantees the data rate requirements of QoS users, but also providea good fairness as well as a higher throughput for BE users. Moreover, the achievedsystem utility is close to the theoretical optimal value.In recent years, cooperative diversity technique has gained much attention as anadvanced emerging transmitting strategy to improve the communication reliability andthe network coverage. In this dissertation, we innovatively introduce the cooperative diversity technique into relay-enhanced OFDMA cellular networks, and investigate theutility-based resource allocation for the scenarios of RSs performing simple subcarriermatching and bit re-allocating respectively, with the consideration of both hard QoS andBE users, and the separate transmission power constraints at both BS and RSs. And theCooperative Joint Subcarrier and Power Allocation (C-JSPA) algorithm and theCooperative Resource Allocation (CoRA) algorithm are proposed respectively, so as tosolve the corresponding relay selection, subcarrier allocation and power controlproblems. Compared with the algorithms without considering cooperative relayingand/or joint resource scheduling, the C-JSPA algorithm has significant advantages inuser utilities and system throughput enhancement, and transmission power saving.Furthermore, due to the resource scheduling at both BS and RSs, the CoRA algorithmcan fully exploit the time, frequency and multiuser diversity of system, and thus furtherimproving the system performance in terms of user utilities, system throughput, powersaving and the number of admitted users.On the other hand, with the explosive growth of wireless data traffic, greenwireless network design has received extensive attention in industry and academia.Therefore, green resource allocation in OFDM relay systems is studied. Consideringboth circuit and transmit power consumption, an energy-efficient resource allocationmodel is proposed with the objective of maximizing the system energy efficiency, so asto jointly optimize the subcarrier-based communication mode selection and powerallocation of source node and relay node. Based on the proposed model, we prove thatgiven the subcarrier-based communication mode selection, the achievable maximumsystem energy efficiency is strictly quasiconcave in total transmission power; and theGreen Power Allocation (GreenPA) algorithm is devised to achieve the energy-efficientpower allocation. Compared with the spectrum-efficient power allocation algorithms,the GreenPA algorithm improves the system energy efficiency significantly.In addition, while RSs are deployed to improve the quality of service of thecell-edge users in one cell, they also bring the interference to the users in the adjacentcells at the same time; and the interference environment is more complex than that ofthe conventional cellular network. Therefore, the inter-cell interference management isanother essential issue of radio resource allocation in relay-enhanced cellular networks.In this dissertation, we focus on a relay-enhanced multicell network with universal frequency reuse, and investigate power control, and joint scheduling, relay selection andpower control for the sum throughput maximization problem respectively, with theconsideration interference and noise impairments, and individual power constraints atboth BSs and RSs. Firstly, the Energy Efficiency based Power Allocation (EEPA)algorithm and the Quasi-Distributed Power Allocation (QDPA) algorithm are proposedfor the scenario where all the cells adopt two-hop relaying communication. The EEPAalgorithm provides an energy-efficient power control solution. The QDPA algorithm isquite attractive for practice systems due to its quasi-distributed implementation and lowcomputational complexity. Numerical results demonstrate that the EEPA and QDPAalgorithms approach the optimal power allocation and the system performance can besignificantly improved in terms of network throughput and energy efficiency.Furthermore, an interference-aware joint resource allocation scheme is studied for thescenario with the coexistence of direct and relaying communication. Specifically, theInter-Cell Interference-based Scheduling (ICIS) algorithm is proposed to address theuser and relay pairing issue. In addition, the Signomial programming-based OptimalPower Allocation (SOPA) algorithm and the suboptimal Minimal Energy-consumptionPower Allocation (MEPA) algorithm are devised as well. Numerical results indicate thatthe “ICIS+MEPA” and “ICIS+SOPA” algorithms approach the optimal resourceallocation, so that the system performance can be significantly improved in terms ofnetwork throughput and energy efficiency.