Akademik Veri Yönetim Sistemi
Tezin Türü: Yüksek Lisans
Tezin Yürütüldüğü Kurum: İstanbul Medipol Üniversitesi, Fen Bilimleri Enstitüsü, Türkiye
Tezin Onay Tarihi: 2024
Tezin Dili: İngilizce
Öğrenci: MUHAMMET FURKAN ÖZER
Asıl Danışman (Eş Danışmanlı Tezler İçin): Hüseyin Arslan
Eş Danışman: Ahmet Yazar
Özet:
Non-terrestrial networks (NTN) have gained significant ground with the advent of 5G and beyond (5GB) wireless communications technologies. Utilization of NTN systems presents strong benefits under various scenarios, owing to the advantages compared to terrestrial networks (TN). To meet different communications requirements from a human-centricity perspective, exploiting the advantages of TN in harmony with NTN is one of the essential topics in 5GB networks. Moreover, machine learning (ML) algorithms have gained special attention lately in satellite communication (SatCom) due to their complex and mathematically intractable nature. In this thesis, a single and multi-parameter-based link selection method is proposed which decides the best connection for the ground users between TN and NTN. Furthermore, an ML-based link selection method is investigated by using the outcomes of the single and multi-parameter-based link selection. The results show that ML techniques give a similar performance. To exploit the usability of ML in different realms, we propose a system framework that utilizes four different management methods for NTN systems and heterogeneous networks (HetNet). The first method is designed to determine the necessity of NTN usage for a specific region. The second method makes a choice between TN and NTN links under a multi-connectivity scenario. The third method tries to choose the appropriate NTN platform for each user. The last method decides which NTN platform is the most reasonable one to meet the NTN usage necessities of the users in the same region. By consecutively employing these methods, cellular communications operators can efficiently manage NTN systems that are integrated into TN infrastructure. For the proposed system framework, new synthetic datasets are generated including region-based information and user-based requirements. It is assumed that there will be many available interconnected sensor systems in smart city networks with the 6th generation (6G). The results indicate that the proposed methods and approaches can be effectively employed to bring benefits to NTN systems in the era of 5GB.