Research On Intelligent Coverage Enhancement Technology For Ultra-Dense Cellular Networks

With the popularity of smart mobile terminals and the rise of various new applications,traditional mobile communication networks have been unable to cope with the rapid growth of business traffic and the increase in user service demand.With the rise of new industries such as the Internet of Things,big data and cloud computing,it is expected that in the 5G era,the business volume of mobile communications will continue to grow exponentially,especially in hotspots,which may even exceed 1000 times than 4G.The amount has dramatically increased the burden on the access network.At the same time,with the rapid development of the mobile Internet,mobile communication services have mainly shifted to indoors,Currently,more than 70% of services data occur in indoor scenarios.Although operators can basically meet coverage after a large-scale 4G network construction in the early stage,the problem of weak coverage of high-rise buildings is serious.In the 5G era,the higher frequency bands are used and the penetration is weaker,so that indoor coverage issues will become a greater challenge.The ultra-dense network form is a typical scenario of future wireless networks,which can effectively solve the problem of difficult indoor coverage.Deploying UDN networks in various indoor scenarios,such as transportation hubs,residential quarters,shopping malls,and office buildings,will greatly ease the traffic pressure on regional networks and simultaneously solve the problem of weak indoor coverage.However,UDN also brings new challenges while improving network performance.Due to the dense distribution of UDN transmission nodes,some common problems in traditional wireless communication networks such as interference coordination and vertical handover are brought about by the deployment of millimeter waves.How to solve these problems will be the key to restrict the performance improvement of UDN.Interference is a problem that has to be solved by UDN.The density of nodes in ultra-dense networks is much greater than the density of active users.Due to the small coverage of nodes,the distance between users and SBS is very close,and there will be multiple SBSs near each user to serve them.In addition,the random deployment of cell locations in the UDN network and the irregular shape of the cells make co-channel interference in the UDN very serious.Due to the high density of base stations and users,and the random deployment of base station and user locations,it is difficult to share data or instantaneous channel information between base stations for complex interference management [26].Therefore,UDN needs a more efficient interference coordination algorithm.Due to the 5G full-band networking,the nodes of the ultra-dense cellular network may work in multiple frequency bands.In ultra-dense heterogeneous cellular networks using millimeter-wave technology,when user equipment moves or there are moving obstacles,the blocking effect may cause the millimeter-wave line-of-sight link to be interrupted [25].In order to avoid disconnection due to blocking effects and ensure seamless network coverage,users should be able to switch freely between millimeterwave nodes and traditional cellular nodes.Therefore,high-quality handover technology is required to provide better service and optimal connection for users.For ultra-dense heterogeneous networks,the interference coordination problem is investigated in this thesis.Considering that under the traditional network architecture,the cost of cooperation between base stations is very large.In order to facilitate the mutual cooperation between base stations,we adopt the C-RAN architecture and implement a centralized algorithm through a cloud resource pool.When adjusting the transmit power of a base station,after the transmit power of a user with a better SINR reaches a certain value,increasing the power will not increase the total capacity of the system.On the contrary,reducing the transmit power of these higher SINR micro base stations can reduce In order to ensure the QoS of edge users,the interference of other micro base stations achieves the overall optimal system throughput.We use a utility function based on user throughput as the basis for base station power allocation.Since the power allocation problem is an NP-Hard problem,the traditional method cannot find the optimal solution in a short time.We use reinforcement learning theory to model the power allocation problem of the base station as a Markov decision problem,and use the Actor-Critic algorithm to obtain the long-term optimal strategy in a dynamic environment.For millimeter-wave heterogeneous cellular networks,the vertical handoff problem is investigated in this thesis.Our goal is to maximize long-term system throughput based on characteristics of user behavior.The characteristics of fast millimeter wave fading,poor penetration,and strong directivity make it easy for obstacles in the indoor environment and user movement to affect them.Therefore,user behavior characteristics need to be considered when solving handover problems.The millimeter-wave base station uses beamforming technology to form multiple directional beams to serve multiple millimeter-wave users at the same time,and the transmitted beams may overlap.Handover in the millimeter-wave network needs to consider the joint selection of millimeter-wave base stations and beams.When the base station performs user handover,it is necessary to make a judgment on the handover timing and an access point according to the dynamic change of the environment,and perform handover in advance to the location where the interruption is likely to occur.By modeling the vertical switching problem as a Markov decision problem and based on reinforcement learning theory,we propose an AC algorithm based on counts.Numerical simulation results show that compared with other traditional switching strategies,our proposed scheme Can effectively improve system throughput and reduce switching overhead.