Estudio del handover en redes lte mediante mediciones de campo: propuesta de un algoritmo básico de predicción para el proceso de handover.

Abstract

In the document, radiofrequency parameters measurements are made in the San Juan sector from Quito. The data collected through the measurement tools compatible with Android: CellMapper, G-NetTrack, and Net Monitor Lite, are analyzed and processed in order to propose a basic algorithm to predict with some level of probability towards the cell to which it is directed the user terminal. The deduced basic algorithm is evaluated with the same data collected, and its effectiveness is determined based on the results. It must be taken into account that the ability of the algorithm to guarantee that handovers are executed depends on the amount of data that can be collected and the capabilities of the measurement applications. For the selection of the handover algorithm, many metrics must be taken into account, such as network conditions, mobile terminal conditions, and weather conditions, since the handover process when using a wireless channel is subject to phenomena, such as ping-pong. And glitches, such as failed handovers, which degrade communication quality. The Machine Learning supervised learning technique is used since this type of technique is applied to obtain a predictive model. This approach becomes an effective solution to improve cellular cell handover performance and increase network utilization.