Research on Cell Sleep Mechanism Based on Clustering and Load Prediction in Ultra-Dense Networks

Abstract

An essential technical means to achieve network expansion in the 5G is to increase the density of small cells, that is, ultra-density network technology. At present, the switch status of cell base stations in ultra-density cells is adaptively adjusted, and cell sleep has become a research hotspot. However, with the rapid rise of the traffic in the network, the load volatility of the cell is also increasing, which is specifically reflected in the variation of the load of the cell with the change of region and time. In view of the high complexity and low energy-saving effect of existing cell sleep technical solutions, this paper proposes a cell sleep mechanism based on clustering and load prediction. The solution first clusters the cells in the current scene, then for the cells in the cluster, the load prediction scheme is used to determine the dormant cells, so as to ensure throughput with lower complexity and achieve the purpose of energy saving. This paper builds a system-level simulation platform for LTE (Long Term Evolution) ultra-density network. The simulation results show that the cell sleep algorithm based on clustering and load prediction can ensure user experience and save energy consumption while expanding the network capacity, thereby providing 5G mobility management and effective technical support for network expansion.