Maximizing deep learning-based energy efficiency in 5G downlink MIMO-NOMA systems by using MLP-CNN.
Multiple input multiple outputs-Nonorthogonal multiple access (MIMO-NOMA), presenting a potential technology to improve system performance and energy efficiency. Nevertheless, the system's effectiveness is hampered by the impact of swiftly changing channel conditions and intricate spatial struc...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Auricle Global Society of Education and Research
2024
|
| Online Access: | http://psasir.upm.edu.my/id/eprint/120012/ http://psasir.upm.edu.my/id/eprint/120012/1/120012.pdf |
| Summary: | Multiple input multiple outputs-Nonorthogonal multiple access (MIMO-NOMA), presenting a potential technology to improve system performance and energy efficiency. Nevertheless, the system's effectiveness is hampered by the impact of swiftly changing channel conditions and intricate spatial structures, restricting its broader applicability. Deep learning plays a crucial role by helping MIMO-NOMA overcome challenges, improve Energy efficiency, and increase capacity and overall system performance in wireless communication networks. This research paper proposes a deep learning-based Multilayer Perceptron-Convolution neural network (MLP-CNN) framework. The framework optimizes the data rate and energy efficiency by addressing the power allocation problems. It can be utilized with multiple convolutional and hidden layers, trained using specific algorithms to solve power allocation problems. Simulation results demonstrate that the proposed framework improves power allocation, overall data rates, and Energy efficiency by around 15% compared to traditional deep neural network (DNN) algorithms, methods and strategies. |
|---|