Research On Vertical Handover And Traffic Offloading Technology In Ultra Dense Network

With the continuous breakthrough and innovation of the new generation of network technology,communication technology is developing in the direction of digitalization,networking and intelligence.Along with the continuous emergence of personal intelligent terminal applications,users have higher and higher requirements on network service quality and request delay.However,the upgrading speed of the existing network foundation cannot keep up with the increase of users' demand for communication services and the explosion of network data traffic.Intensive network deployment is one of the effective measures to expand network capacity and support reliable data services.Ultradense heterogeneous network deployment adopts the method of dense deployment of micro-base station to expand the network coverage and improve the network capacity.Therefore,in order to cope with rising web service request,user' personalization attribute and communications business diversity,mobility management technology and traffic offloading technology can be used as safeguards to obtain properly effective network services under super dense heterogeneous cellular network deployment.Firstly,a vertical handover mechanism of business type classification is proposed to solve the mobility management problems such as frequent handover,ping-pong handover and high call drop rate of mobile users in ultra-dense heterogeneous networks.In this thesis,in order to improve the network throughput and the quality of user's service experience,it considers to use the type of communication business that terminals use.Through deep flow inspection technology to analyze the behavior characteristics of accessed network packets,then the classification of service business is obtained.And in the process of handover,classification of service business is used to make dynamic handover come true,which could avoid unnecessary handover and reduce network signaling load.Simulation results show that,compared with the existing vertical handover algorithm based on received signal strength,the number of ping-pong handover is reduced,the probability of handover failure is decreased,and the throughput of the system is improved.Secondly,in this thesis,a deep learning-based traffic offloading mechanism is proposed to solve the problem of unbalanced load on the core network that due to the explosion of mobile data flow,which affects the energy efficiency of the system.In this thesis,the design of users' terminal traffic offload system according to existing study is described in detail.The offloading mechanism describes how to make the offloading resources allocation of users' packet traffic.Taking the users' energy loss and networks' energy loss as the cost of traffic offload.Then the model is transformed into a dynamic optimization problem according to markov process theory,and the deep Q network is used to solve the optimization question.Simulation results show that the proposed algorithm can effectively reduce network load and improve system energy efficiency.