An Intelligent eNodeB Architecture for Adaptive Mobility Load Balancing in Long Term Evolution Networks

ABSTRACT:

4G LTE network architecture revolutionized mobile communication in so many ways. Typical of which includes, changes to its overall architecture, support for flexibility to its eNodeB functions and self-organizing operations. Self-organizing networks (SON) has become a key feature in modern mobile communication architecture by reducing perational and capital expenses through automation. Despite the improvement, 4G LTE networks offers and its foundational role in 5G networks the network still suffers from the congestion problem. Many researchers have employed machine-learning techniques (MLT) to reduce the congestion problem through the mobility load-balancing scheme (MLB) with lots of encouraging results. However, the architecture of the eNodeB is limiting these gains. We propose an intelligent eNodeB architecture that can help the machine learning algorithms had better infer the dynamics surrounding the eNodeBs on the network. We also introduced for consideration the handon parameter, a functional inverse of handoff, a parameter that serves as a link to the mobility load balancing dynamics around the eNodeB. We adopted the multi-layer perceptron neural network (MLPNN) to the adaptive mobility load balancing (AMLB) architecture. Simulation results showed an encouraging performance against the traditional MLB (TMLB) and the without MLB (WMLB) schemes. The AMLB witnessed an improvement in the unsatisfied and blocked requests over the TMLB and the WMLB

Keywords: Adaptive Mobility Load Balance, Artificial Neural Network, eNodeB, Handon, Long Term Evolution.