Cell Selection Mechanism Based on Q-learning Environment in Femtocell LTE-A Networks

Ammar Abdulrazzak Bathich (Al-Madinah International University)
Saiful Izwan Suliman (Faculty of Electrical Engineering, Universiti Teknologi MARA (UiTM), Jalan Ilmu 1/1, Shah Alam 40450, Selangor,)
Hj. Mohd Asri Hj. Mansor (Faculty of Electrical Engineering, Universiti Teknologi MARA (UiTM), Jalan Ilmu 1/1, Shah Alam 40450, Selangor,)
Sinan Ghassan Abid Ali (Faculty of Computer Technologies Engineering, Iraq University College Al-Estiqlal St, Basrah)
Raed Abdulla (School of Engineering, Asia Pacific University of Technology and Innovation (APU) Jalan Teknologi 5, Kuala Lumpur 57000, Malaysia)

Journal of ICT Research and Applications
Vol. 15 No. 1 (2021)

Universal mobile networks require enhanced capability and appropriate quality of service (QoS) and experience (QoE). To achieve this, Long Term Evolution (LTE) system operators have intensively deployed femtocells (HeNBs) along with macrocells (eNBs) to offer user equipment (UE) with optimal capacity coverage and best quality of service. To achieve the requirement of QoS in the handover stage among macrocells and femtocells we need a seamless cell selection mechanism. Cell selection requirements are considered a difficult task in femtocell-based networks and effective cell selection procedures are essential to reduce the ping-pong phenomenon and to minimize needless handovers. In this study, we propose a seamless cell selection scheme for macrocell-femtocell LTE systems, based on the Q-learning environment. A novel cell selection mechanism is proposed for high-density femtocell network topologies to evaluate the target base station in the handover stage. We used the LTE-Sim simulator to implement and evaluate the cell selection procedures. The simulation results were encouraging: a decrease in the control signaling rate and packet loss ratio were observed and at the same time the system throughput was increased.