Power allocation for cooperative NOMA systems based on adaptive-neuro fuzzy inference system

dc.contributor.advisor Çırpan, Hakan Ali
dc.contributor.author Üçbaş, Melike Nur
dc.contributor.authorID 504191352
dc.contributor.department Telecommunications Engineering Programme
dc.date.accessioned 2025-05-21T12:16:57Z
dc.date.available 2025-05-21T12:16:57Z
dc.date.issued 2023
dc.description Thesis (M.Sc.) -- Istanbul Technical University, Graduate School, 2023
dc.description.abstract Innovative technologies improving capacity, coverage, energy efficiency, and service quality are required to meet the exponentially increasing traffic demands in wireless communication systems. Non-Orthogonal Multiple Access (NOMA), which allows multiple users to transmit their data simultaneously at the same frequency and time interval, is a promising radio access technology to cope with the challenging requirements of 5G and beyond systems. However, the importance of energy efficiency in cellular networks for the NOMA becomes a major issue as the number of users increases. In a cooperative NOMA architecture, relays are effective in increasing system performance and reducing outage probability. The power allocation in a cooperative NOMA system is a challenging task having a significant impact on the user's perceived quality of service. In this thesis, a fuzzy logic (FL) based relay selection and power allocation approach are proposed for a multi-relay NOMA system with imperfect successive interference cancellation. The power is allocated between the NOMA user pair within a resource block in such a way that the rate fairness is maximized and the system outage is minimized. In order to demonstrate the effectiveness of the proposed system model, we utilize a network scenario including a base station, a variable number of relays, and two users. Relay selection and power allocation are performed using two different fuzzy inference systems (FIS). These FISs are created by training parameters such as channel coefficients, signal-to-noise ratio (SNR), and interference with the Adaptive-Neuro Fuzzy Inference System (ANFIS) method. The first FIS is designed to find relays that can achieve the minimum rate required for communication between the base station and relays. Its input parameters include the channel coefficient, SNR, self-interference, and residual interference between the base station and relays. The output of the FIS is the minimum achievable rate for the users. The second FIS is applied only for the relays that satisfy the minimum data rate requirements. The objective of the second system is to distribute the power fairly between the users. The input parameters of the second FIS are the channel coefficients, SNR, and residual interference between users and relays. The power allocation coefficient for a strong user is obtained as an output of the second FIS. The numerical results obtained by FL are close to the optimum outage probabilities and rate fairness results for all experiments when the number of relays and SNRs are varied. The computationally effective FL may be successfully applied at run time for the power allocation in a cooperative NOMA system, which gives rise to promising outcomes.
dc.description.degree M.sc.
dc.identifier.uri http://hdl.handle.net/11527/27129
dc.language.iso en
dc.publisher Graduate School
dc.sdg.type Goal 9: Industry, Innovation and Infrastructure
dc.subject Power allocation
dc.subject NOMA systems
dc.subject Innovative technologies
dc.title Power allocation for cooperative NOMA systems based on adaptive-neuro fuzzy inference system
dc.title.alternative Uyarlanabilir nöro bulanık çıkarım sistemine dayalı işbirlikli NOMA sistemleri için güç tahsisi
dc.type Master Thesis
Dosyalar
Orijinal seri
Şimdi gösteriliyor 1 - 1 / 1
thumbnail.default.alt
Ad:
807157.pdf
Boyut:
2.06 MB
Format:
Adobe Portable Document Format
Açıklama
Lisanslı seri
Şimdi gösteriliyor 1 - 1 / 1
thumbnail.default.placeholder
Ad:
license.txt
Boyut:
1.58 KB
Format:
Item-specific license agreed upon to submission
Açıklama