HsMM-based mobility aware cell association method for dynamic bandwidth management in 5G-FRANs

Abstract

In this paper, a cell association method for proactively managing bandwidth in 5G fog radio access networks 5G FRANs) by considering user mobility is developed. The proposed method jointly utilises cell identifications (cell-IDs) information and distributions of sojourn time to develop a hidden semi-Markov model (HsMM) for predicting the most appropriate next cell that satisfies downlink rate requirements. Furthermore, the cell that offers the optimal rate is determined based on an optimization solution. The proposed HsMM considers ping-pong effects, missing values, and limited context from wireless networks to understand user mobility patterns. The result shows that the HsMM-based scheme yields a high accuracy performance according to the number of observation time intervals. Also, the proposed scheme produces an improved user satisfactory percentage compared to the existing schemes.