Hücreler Arasi Geçiş Şemalarinin Performansi Üzerine Deneysel Bir Çalişma

Handover (HO) is a key process for network management in terms of uninterrupted data transmission. In cases where the time-to-trigger (TTT), one of the HO parameters, is not determined correctly, it causes ping-pong effects or radio link failure. Past HO events can be used to make HO decisions earlier, while avoiding the ping-pong effects in HO operations. In the literature, it has been mentioned that ML algorithms can enhance network mobility. In this study, based on past HO data, neighbor cells are classified according to target cell likelihood using supervised machine learning algorithms (SML). The data set that is obtained regardless of the direction and speed of the user is used. Signal quality measurements were taken with the G-Net Track Pro mobile app. Afterwards, it is presented that SML can be applied to different schemes created for TTT initiation and HO operations can be performed by determining the TTT start time earlier by comparing the accuracy rates.