Research On Clustering-based Energy Consumption Management In Ultra-dense Networks

The Ultra-Dense Networks(UDN)can shorten the access distance of users,improve the transmission quality of links,and increase the reuse of spectrum resources through large-scale deployment of wireless access points,which can improve the throughput of users and meet system capacity requirements.However,in UDN,the randomness,densification and heterogeneity of node distribution lead to more complex and changeable network topology and stronger energy consumption and interference,which further results in more difficult in the energy consumption management,interference management and mobility management.Therefore,this thesis focuses on the single-layer and multi-layer UDN scenarios,takes energy consumption management as the research object,clusters the base stations of the same layer and different layers,and formulates the energy consumption management strategy based on the base station on/off technology,so as to reduce the network energy consumption and improve the network energy efficiency.The main research work of this thesis is summarized as follows:(1)A base station on/off energy consumption management scheme based on density clustering is proposed.First,in view of the large-scale random distribution of base stations and the complex network topology in the single-layer UDN,the local density of base stations is analyzed to cluster the base stations in the same layer according to the distribution characteristics of base stations,and the effective partition management of the whole network and the logical simplification of the network topology are realized through the clustering of base stations.Then,for the calculation complexity of base station on/off and the change of network load caused by it,the load change caused by base station on/off is measured by the interference between base stations,and the combination problem of base station on/off is transformed into a selection problem to reduce the search space of base station on/off,so as to reduce the complexity.Finally,the simulation results show that the proposed scheme can effectively improve energy efficiency.(2)A base station on/off energy consumption management scheme based on dynamic affinity propagation clustering is proposed.Aiming at the coverage mode,load conditions and energy consumption differences of multi-type base stations in multi-layer UDN,a base station on/off energy consumption management scheme based on dynamic affinity propagation clustering is proposed.First,considering the handover cost of extra energy consumption due to the change of user association status caused by the base station on/off,a cell selection strategy for reducing handover cost is proposed.Then,in view of the differences in location distribution,coverage and transmission power of base stations of different layers,and in order to improve the cooperation of base stations of different layers,a dynamic affinity propagation clustering algorithm is proposed.By using the degree of interference between base stations as a similarity measure,base stations of different layers are dynamically divided into base station clusters,and base stations of different layers in the same cluster perform on/off cooperation.Finally,based on the clustering results,a base station on/off strategy based on interference similarity is implemented in each base station cluster to reduce network energy waste caused by non-uniform time-varying traffic loads.Simulation results show that the proposed scheme can effectively improve the network energy efficiency and reduce the switching cost.