Machine Learning Based Classification Algorithm for AP Selection in Cell-Free MIMO Systems

Abstract

With the rapid development of technology, cellular networks in wireless networks are insufficient to meet the demands. In order to provide a correct and good service to each user, communication systems must change. Although cell-free networks have many advantages over cellular networks, since there are too many users and access points (APs) in cell-free networks, AP selection is very important. In this thesis, the AP selection model has been studied and compared five different machine learning classification methods. The campus of Izmir Katip Celebi University has been chosen as the environment where the study will be carried out, and capacity values have been obtained from the users and APs that have been placed on the campus in the simulation environment. Numerical calculation results have been obtained from the Wireless Insite (WI) software. The AP selection to be created with the capacity values has been supported by artificial intelligence algorithm techniques. With two different data sets have been compared, better results have been tried to be obtained with different feature values. As a result of the comparisons made, the best machine learning classification technique used has been proposed.