.png){kind=icon}
.png){kind=icon}
Return
A Supervised Learning-Assisted Partitioning Solution for RIS-Aided NOMA Systems
.png){kind=icon}
PRE
AI
DOI：10.1109/TCCN.2024.3373637.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Thanks to its capacity for producing intelligent radio environments that are both efficient and affordable, reconfigurable intelligent surfaces technology is gaining recognition as a potential solution for advanced communication systems. Efficient information processing is crucial for smart surfaces to effectively respond to electromagnetic signals, however achieving this requires additional resources such as computing time, storage, energy, and bandwidth. To address these challenges, model-agnostic methods such as machine learning can be an effective solution, as ML employs trainable variables to examine raw data and generate valuable outcomes. This study introduces a novel approach that integrates a hybrid RIS and utilizes an uplink non-orthogonal multiple access transmission from the users to the base-station. The proposed scheme utilizes supervised learning for RIS partitioning to optimize RIS element distribution that minimizes interference between users situated in the RIS's non-line-of-sight. The proposed system achieves similar achievable rates and fairness among users as the current advanced iterative algorithm described in existing literature, while significantly reducing the time and complexity involved. A theoretical outage probability formulation is derived along with computer simulations and comparisons presented to assess system outage and bit error probability results for varying quality-of-service conditions and successive interference cancellation scenarios.
Keywords:
NOMA
Wireless communication
Complexity theory
Reconfigurable intelligent surfaces
Quality of service
Resource management
Interference cancellation
Machine learning
supervised learning
reconfigurable intelligent surfaces
non-orthogonal multiple access
RIS partitioning
Journal
I
IF:
7
Papers:
1.5K
Citations:
5.5K