A Markov-Based Prediction Algorithm for User Mobility at Heterogeneous Cloud Radio Access Network

To provide high-quality services to the explosion of users in cellular networks, forecasts of users' next location have been studied. But it is difficult to predict users' next location in urban areas because of mobility and dynamic road and traffic condition. To address the problems, it is used that users tend to move based on social relations. If the user's next location is predictable, the traffic changes will be also predictable. It can be applied to various studies, such as base station switching and handover forecasts that can solve energy consumption and service delay problems. In this study, we propose a Markov-based prediction algorithm to forecast the next location of users in H-CRAN (heterogeneous cloud radio access network). We use RRH(radio remote head) trajectory that consists of histories stored serving RRH for each user. Simulation results and discussions demonstrate the prediction accuracy compared with those of autoregressive models.