Research On Base Station Sleep Mode Technology Under 5G Dense Heterogeneous Network

The fifth generation of communication technology(5G)is a new generation of mobile communication system that is developed to meet the demand of mobile communication after 2020.Dense heterogeneous network technology is one of the key technologies to ensure the explosive traffic growth in the 5G scenario.However,the deployment of a large number of micro base stations could leeds the linear growth of the wireless communication system's energy consumption.In order to improve the energy efficiency of the system and build the green wireless communication system,the base station sleep mode technology has been drawn much attention due to its easy implementation and no modification in hardware.There are some problems in present base station technologies in both academic and industrial,such as high computational complexity and lack of some important information.In order to solve these problems,two novel base station sleep mode technology,which improve the wireless communication system's energy efficiency,are proposed and simulated in this paper.In order to reduce the system energy consumption as much as possible that maintain the Quality of Service(QoS),this paper first proposes a hybrid-distributed base station sleep mode technology.Hybrid-distributed sleep mode technology combines the advantages of centralized and distributed sleep mode technology.The proposed technology is divided into two parts,the cword-awake algorithm and the self-sleep algorithm.A clustering algorithm is used to reduce the computational complexity,cword-awake algorithm with macro-base-station-assisted and neighboring-micro-base-station-assisted is used to solve the problem of laking information of sleeping node,and self-sleep algorithm is used to flexible scheduling.System-level simulation shows that the proposed hybrid-distributed sleep mode technology can reduce energy consumption by 13%averagely,and increase throughput by 9%averagely,compared with no-sleep technology.Secondly,this paper introduces machine learning into the base station sleep mode algorithm,which is used to improve the proposed hybrid-distributed base station sleep mode technology,and proposes a machine-learning-based base station sleep mode algorithm.The artificial neural network is used to train the micro base stations in a cluster.Since the training process can run remotely and the training results are deployed in the micro base stations within the cluster,the computation of the base station sleep mode algorithm can be effectively reduced.In the proposed algorithm in this paper,the base station's traffic load and the direction of arrival vector are taken as the input of artificial neural network,and the sleep state of micro base stations is taken as the output.The simulation results show that the machine-learning-based base station sleep mode technology can reduce the energy consumption by 3.1%averagely,while maintaining the same throughput,compared with the hybrid-distributed base station sleep mode technology.