Mobility Prediction Method to Optimize Resource Allocation in Heterogeneous Networks

With the explosive development of wireless communication technologies and popularization of smart phones, mobile data traffic grows rapidly. The 5G heterogeneous networks constituted by macro cells and femtocells are widely researched as a promising approach to expand the network capability. However, for the cases of moving user equipments (UEs), this flexible and complicated networking method may lead to frequent handover and unsatisfied resource assignment. In this paper, we propose a mobility prediction scheme to optimize the resource allocation in heterogeneous networks based on order-2 Hidden Markov Model (HMM). With the prediction result of the possible next location of UE, the target cell may prepare necessary resource according to the service requirements. The user's historical movement trajectory is clustered into several groups to represent its main activity states. The user's service request preference in these locations are analyzed accordingly. Then, we build the mobility prediction model with order-2 HMM. Corresponding required resources are also analyzed and allocated in advance. The simulation results show that the proposed scheme has a good performance in prediction accuracy, drop rate and resource utility for the heterogeneous networks.