Research On Dynamic Clustering For Multi-Cell Cooperative Reception

With the sharply increasing of wireless data and equipment,the shortage of frequency resources is arising.The aggressive frequency reuse among cells can dramatically improve the frequency efficiency,however,the reuse of the time-frequency resource block among adjacent users induces severe Inter-Cell Interference(ICI),which leads to negative impact on system throughput and Quality-of-Service(QoS).To our knowledge,ICI has emerged as one of the key factors that heavily constrains the capacity,especially in Ultra Dense Networks(UDN).Multi-Cell Processing(MCP)has been emerged as a promising technique in combating ICI and increasing the capacity,where multiple Base Stations(BSs)exchange information through the backhaul for joint transmission/reception.In the uplink of multi-cell systems,cooperative reception,as a typical MCP scheme,can effectively improve the system capacity and the QoS of cell-edge users.Although the full cooperation among all BSs can achieve the largest cooperative gain in theory,it brings the extra enormous overhead,including the information exchange overhead among cooperative BSs,as well as the delay and computation complexity imposed by cooperative reception.Luckily,the partial cooperation among clustered BSs can obtain most of the cooperative gains with extremely low cooperative costs,thus attracts a lot of attention in recent years.An efficient clustering mechanism is a prerequisite for the subsequent MCP,and it is critical to the tradeoff between the cooperative gain and cost,i.e.,the cooperative resource efficiency.This dissertation dedicate to the designing of dynamic user-centric cluster for uplink cooperative reception in multi-cell wireless networks.Taking the trade-off between cooperative gains and costs into consideration,the practical constraints are involved as well.Thus,a series of optimization model of clustering is built,and multiple user-centric dynamic clustering algorithms with various constraints are proposed.It is necessary to highlight that the main contributions of this dissertation are listed below:(1)Considering the minimization of the cooperative cost,an optimization model minimizing the cluster size is proposed.With QoS provisioning for cell-edge users,a Minimizing Cluster Size based dynamic clustering algorithm without Constraints on Cooperative Resources(MCS-W/OCCR)is proposed,where the Lagrangian dual and sub-gradient approach are adopted to obtain the approximate solution.By introducing the upper bounds on the cluster size and maximum cooperation times of each BS,a Minimizing Cluster Size based dynamic clustering algorithm with Constraints on Cooperative Resources(MCS-WCCR)is further proposed.Simulation results show that,the proposed MCS-W/OCCR algorithm achieves best performance in the outage analysis when compared with existing related algorithms.The proposed MCS-WCCR algorithm can guarantee the fairness among BSs with the modest computational complexity,as well as greatly reduce cooperative costs with QoS provisioning.(2)Considering the combinatorial optimization of the cooperative transmission performance and the cooperative cost,the user-centric dynamic clustering problem is modeled as a multi-objective 0-1 Integer Programming(IP)problem,where the minimization of numbers of outage users and the minimization of the cluster size in the network are both taken as the optimization objectives.Giving up the cluster design of outage users,the Minimizing Numbers of Outage Users based dynamic clustering algorithm WithOut Constraints on Cooperative Resource(MNOU-W/OCCR)is proposed and optimal solution is proved.A Minimizing Numbers of Outage Users based dynamic clustering algorithm with Constraints on Cluster Size(MNOU-CCS)is further proposed.Specially,with the extension of the multi-carrier cooperative burden of each BS,the multi-objective 0-1 IP problem is modeled with constraints on cooperative resources.Based on this,the Minimizing Numbers of Outage Users based dynamic clustering algorithm with Constraints on Maximum Cooperative Burden of BS(MNOU-CMCB)is proposed.Compared with related clustering algorithms,including MCS,the proposed MNOU-W/OCCR algorithm can achieve the best outage performance as well as significantly reduced to average size of cluster and computational complexity.With a small loss on average transmission rates,the proposed MNOU-CCS and MNOU-CMCB algorithms can provide largest rate gains and minimize cluster size when consuming a unitary cooperative resource.(3)Considering the holistically optimization of the cooperative resource efficiency,the unitary cooperative gain is defined to evaluate the efficiency.Specifically,the unitary cooperative gain is defined as the rate gain achieved by the unitary cooperative resource,i.e.,one cooperation link occupying the backhaul on one time-frequency resource block.A Maximizing the Unitary Cooperative Gain based dynamic clustering algorithm with Constraints on BS Distance(MNCG-CD)is proposed to control the cluster size and distance among cooperative BSs.Taking the power consumption on BS and the rate for cooperative transmission into consideration,a Maximizing the Unitary Cooperative Gain based dynamic clustering algorithm with Constraints on Cooperative Rates(MNCG-CCR)is finally proposed.Simulation results indicate that,the proposed MNCG-CD algorithm effectively controls the maximum BS distance when approaches to the upper-bound MNOU-CCS algorithm.The proposed MNCG-CCR algorithm exhibits significantly reduce to the maximum BS distance,the maximum cooperation times and the cooperative rates on the backhaul,as well as achieve the near performance of the upper-bound MNOU-CCS algorithm,which indicates the good resource efficiency under effective control of BSs' cooperative burden.