| With the wide use of LTE,more and more people can enjoy the convenience of high-speed mobile network.Although LTE can provide the downlink transmission rate up to 100M,with the people of all kinds of business such as high-speed,high-definition video,mobile phone games and other business needs continue to improve,and cloud computing,widely used in large data,estimated in the next ten years,traffic demand will increase by 1000 times,the current LTE has gradually difficult to meet the increasingly stringent and load demand.Therefore,the study of 5G is in line with the requirements of 5G LTE with higher transmission rate,lower latency.In order to meet the stringent requirements of 5G,a variety of new technologies,such as software de-fined network(SDN),large-scale MIMO,ultra dense network,are proposed.SDN is a new type of network architecture,and the difference between the traditional network structure is SDN data plane and control plane separation,the data plane forwarding equipment become simple equipment,and all of the control logic through the SDN controller in the control plane to complete.Using this framework can effectively solve the problem of traditional network management in complex structure,and can enhance the flexibility of the network.Large scale MIMO benefits mainly reflected in:the depth of mining spatial dimension resources,spatial degrees of freedom and the base station so that multiple users in the network can use large-scale MIMO provides the same time-frequency resources and communication,to increase the density of base station bandwidth and greatly improve the spectrum efficiency without.The beam can be concentrated in a narrow range,thus greatly reducing interference.The transmitting power can be greatly reduced,so as to improve the power ef-ficiency.When the number of antennas is large enough,the most simple linear precoding and optimal linear detector,and the noise and irrelevant interference can be neglected.People put forward a variety of programs and a large number of algorithms for the above technologies.The main work of this paper is to study the implementation of SDN architecture based on mobile network.Firstly,this paper introduces the SDN system architecture and key technologies,such as the southbound interface protocol and SDN controller.After that,the research status of the use of SDN architecture in wireless network is listed.It mainly includes the improvement of architecture,the improvement of access network and related patents.Considering the high power consumption and high load of massive MIMO,it does not meet the needs of the current green communication.In the third chapter,based on the SDN architecture,a method and a device for massive MIMO wireless network dynamic network downlink transmission network are proposed.The method of clustering by the user,the base station transmit antenna selection,resource allocation,adjust the frequency of 3D dynamic user QoS based on space time,the whole network management with SDN controller,on the premise of satisfying customer demand,improve the network energy efficiency,reduce network overhead.Finally,in order to make full use of the advantages of micro base station,improve the micro average capacity of the cell edge users and to achieve the optimal value of CRE,this paper proposes a reinforcement learning and SDN adjustment method based on extended coverage.Q-learning is used to each user to search the optimal value of the CRE offset,and the use of SDN framework can reduce the burden of signaling communication between base station and pico-base station.The simulation results show that the proposed method can reduce the outage probability and improve the average throughput of the base station and the edge user throughput. |
PDF Full Text Request