The Management Research Based On User Mobility Prediction In Heterogeneous Wireless Networks

The development of heterogeneous wireless networks (HetNet) is in line with the trend of wireless network system integration in the future, in which, handover technology under heterogeneous wireless networks is a hot research area. A large number of relevant literatures have studied on it in recent years. It greatly affects the user’s service quality, which relates to the seamlessly connection between different networks. Therefore, the study and implementation of the technology related to heterogeneous wireless networks will have an important significance.In this context, this paper is focus on the study of handover management direction for the heterogeneous wireless network mobility management. Under the study background of handover management, we innovatively study the prediction handover algorithm, which is based on the user mobility. This paper focuses on that how to use prediction handover technology to achieve better accurate handover and reduce unnecessary switching in heterogeneous wireless networks. Meanwhile, the paper also carried out corresponding simulation analysis for different scenarios to confirm the nature of the algorithm.Firstly, the paper describes the background and the existing status of heterogeneous wireless. With mobile data traffic increasing and more and more high-bandwidth requirements for the mobile user, the integration of heterogeneous wireless networks has become a must development and makes sure the user’s good network experience. Then we combine with the practical application of existing heterogeneous wireless networks, further overview and analysis the handover management in heterogeneous wireless networks, and for which, we conclude the existing prediction handover algorithm and do some relative research. With the basis of the existing algorithms through research and analysis, this paper proposes an innovative combination prediction handover algorithm. Due to the heterogeneous network downlink pilot signal/reference signal is a time-series signal, and considering this feature, we decompose the downlink pilot signals, through the use of T-S Fuzzy Neural Network(FNN) and Auto Regressive(AR) model, which is a time series prediction model, to forecast and analysis the data based on user history more accurately. Secondly, combining the feature of different user velocity characteristics under different scenarios in heterogeneous network, we analysis the algorithm separately for different scenarios. And the simulation results confirm the effectiveness of this algorithm.In addition, in order to demonstrate the effectiveness of the algorithm more accurately, in the simulation process, we compare this algorithm with prediction handover algorithm based solely on the autoregressive model. By comparison we could prove that, for the simulation index, which mainly with the index of user handover frequency and user cumulative function, the simulation results show the simulated targets are superior than contrast algorithms under different scenarios.