Research Of Handoff Optimization Algorithm Based On User Mobility Prediction

With the rapid development of communication technologies and the widespread popularity of intelligent terminals,mobile data traffic shows explosive growth.The transmission demands of high-speed,low latency and others are also increasing.This trend puts load pressure for the network,and has a higher demand for the reasonable resources allocation.The emergence of hierarchical heterogeneous networks solves the problem of poor coverage and lack of bandwidth resources in the indoor and crowded areas.However,due to the user's movement,the terminal device will switch between different cells frequently and the network resources need to be re-allocated.How to solve the above problem and realize proper resource allocation and smooth handover for the moving user is an important challenge for the mobile communication network.At present,the mobility prediction technology can be used as an effective solution to the above problems.By analyzing the moving characteristics of the user,the relevant mobility model is established,and the next moving area of the user is predicted.Then the appropriate handover target base station is selected in advance using the prediction result,thereby reducing the number of unnecessary handovers.At the same time,the movement trend of the group can provide a reference for the rational allocation of network resources and improve the resource utilization efficiency.However,due to the complexity of the user movement in a particular scene,the mobile behavior of users is hard to model precisely.Second,the increase in the order of moving data puts forward higher requirements for the ability of analyzing data.Therefore,in a particular scenario,it is interesting to study the user behavior for mobility modeling and to develop a reasonable forecasting scheme.In view of the above problems,this paper considers two typical scenarios,indoor and outdoor,and analyzes the mobility prediction of users for better handover optimization.The main work is as follows:First,a mobile forecasting scheme based on Hidden Markov Model(HMM)is proposed to optimize the switching process for indoor office scenes.In the HMM model,the location of the home base station is defined as a hidden variable state,and the user's movement trajectory is defined as an observation sequence.The prediction time is divided into several time periods according to the working schedule of the users in the scenario,and the state transition probability matrix is calculated respectively.In addition,considering the influence of walls and obstructions on the signal in the indoor environment,we introduce the wall influence factor in the medium and low signal intensity regions to adjust the observation probability matrix,and consider the building topology to eliminate the non-existent prediction path.Finally,the effectiveness of the mobility scheme is proved by simulation.The scheme improves the accuracy of indoor user motion prediction and reduces the number of unnecessary handover between cells.Secondly,a mobile forecasting scheme based on data mining technology is proposed to optimize the handover process.A large number of location data are processed quickly and the hidden movement pattern is hidden.By analyzing the mobile characteristics of the users in this scenario,the Kentucky campus is divided into several forecasting regions according to the distribution of the users on the campus,and the forecasting time is divided into different time periods according to the class and the schedule.Then,the user movement trajectory of the region is divided into several mobile user groups by fuzzy C-means clustering method for a specific period of time in each prediction area.Finally,the sequence pattern mining technique is used to find out the frequent trajectories of each user's moving trajectory,and the irregular moving trajectories are excluded.By matching the new user's historical trajectory and the moving pattern of the corresponding user group,the sequence pattern with the highest matching degree is selected to predict the user's movement at the next moment of the location area.And the optimal switching target base station is decided according to the predicted position information,thereby optimizing the switching process and ensuring the communication quality of the user.