A Machine Learning Approach for Handover in LTE Networks with Signal Obstructions

Abstract

Legacy cell-deployment strategies have been adapted to fulfill the increasing demand for wireless broadband internet access. One of them, the Hierarchical Cell Structure (HCS), that is already in use in LTE-A and it is considered essential for the 5G, consists of the deployment of several types of small cells under the umbrella of macrocells, creating overlaid coverages. Due to their low power and bellow-rooftop-level, sometimes indoor base stations, the small cells are severely affected by the surrounding obstacles, making the perceived Quality of Service (QoS) of the users subject to fast variations, thus rendering ineffective the classical approaches to mobility management, that are unable to predict those severe fading situations (coverage holes). Considering the amount of available information on the network performance and the evolution of real-time processing capabilities, the enhancement of LTE functionalities (such as the handover) by means of machine learning algorithms became possible. This work proposes and evaluates the performance of a machine learning based approach to handover in scenarios with the presence of signal-blocking obstacles. We use the ns-3 simulation for our proof of concept simulations. Our machines learn from experience and they are, therefore, able to choose the eNB that will most likely offer the user the highest long term QoS after the handover procedure, even in severe propagation conditions. The proposed schemes substantially improve the users' QoS in certain circumstances.