Clustering Based Interference Management Research On Dense Dynamic TDD Networks

With the flourish development of mobile internet,various emerging multi-media traffic started growing explosively,which poses a great challenge to the transmission capacity of the network.By densely deploying low-power micro-cellular base stations,the spatial reuse rate of spectrum resources can be increased,and the network capacity can be greatly improved.At the same time,due to people's different business requirements,the traffic in cellular networks show the characteristics of sudden and dynamic,resulting in the asymmetry of traffic distribution in space and time.Dynamic Time Division Duplex(DTDD)technology allows the base stations to adjust cell-specific UL/DL subframe configurations dynamically according to the instantaneous traffic conditions in small cells,which can support the asymmetric internet traffic better,and further improve the utilization of spectrum resources.Dense dynamic TDD network will be an important direction of cellular network development in the future.However,the application of dynamic TDD also brings crosssubframe interference,in which the interference of downlink base station to uplink base station will seriously affect the performance of UL transmission,and weaken the performance gains brought by dynamic TDD.Therefore,it is necessary to study the interference management methods which are suitable for dense dynamic TDD networks to reduce the cross-frame interference between small cells,and further improve the network performance.In order to reduce the strong cross-subframe interference among different small cells which are set different subframe configurations,this paper studies the cross-subframe interference cancellation technology based on cell clustering.By dynamically cell clustering process,base stations in the same cluster use the same subframe configuration to eliminate the crosssubframe interference.At the same time,different clusters are allowed to set different subframe configurations to improve the utilization of network resources.Considering the difference of interference intensity and traffic distribution among different cells,we first propose a cell clustering criterion based on interference and traffic difference,which can cluster base stations dynamically according to the clustering criterion to maximize the overall throughput within the cluster.Then,we choose some users to re-associate the base station in the cluster,which reduces the traffic difference between cells in the same cluster,and improves the matching rate between the traffic distribution of small cells and the unified subframe configuration in cluster,and further improves the utilization of network resources.Finally,the tradeoff between the overhead of the dynamic clustering algorithm and the network performance is studied.By constructing utility function,the optimal clustering period for the balance between the overhead of the clustering algorithm and the network performance is determined.The interference management method based on clustering designed in our paper provides a theoretical and experimental basis for the actual deployment of dense dynamic TDD networks.