Mobile Power Allocation Intelligent Optimization Algorithm for Cooperative NOMA Network Based on CBAM-BiLSTM

The non-orthogonal multiple access (NOMA) technology can greatly improve the spectral efficiency of wireless communication systems. The incorporation of NOMA technology into a 5G mobile communication network has the potential to significantly improve communication performance. First, we establish an...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE transactions on vehicular technology Jg. 73; H. 5; S. 7131 - 7139
Hauptverfasser: Xu, Lingwei, Cao, Shubo, Fu, Xingyue, Qin, Xujiang, Li, Xingwang, Gulliver, T. Aaron
Format: Journal Article
Sprache:Englisch
Veröffentlicht: New York IEEE 01.05.2024
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Schlagworte:
5G mobile communication
Algorithms
CBAM-BiLSTM
Communication
Feature extraction
Feature maps
intelligent power allocation
Mobile wireless communication
NOMA
non-orthogonal multiple access (NOMA)
Nonorthogonal multiple access
Optimization
Optimization algorithms
Performance prediction
Prediction algorithms
Relays
Resource management
Signal to noise ratio
Wireless communication systems
ISSN:0018-9545, 1939-9359
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The non-orthogonal multiple access (NOMA) technology can greatly improve the spectral efficiency of wireless communication systems. The incorporation of NOMA technology into a 5G mobile communication network has the potential to significantly improve communication performance. First, we establish an mobile cooperative NOMA multi-user network. The exact outage probability (OP) expressions are then derived, and the effect of the power allocation on OP performance is investigated. Finally, we design a CBAM-BiLSTM network and propose an intelligent power allocation optimization algorithm based on system efficiency and user fairness. The CBAM-BiLSTM network is a structure based on convolutional block attention mechanism (CBAM) and bidirectional long short term memory network (BiLSTM). CBAM performs feature selection and weights the spatial and channel dimensions of feature maps to improve the network's classification accuracy. BiLSTM can fully utilize contextual information and handle long-term dependencies, thereby providing more comprehensive and accurate modeling capabilities and predictive performance. Simulation results indicate that, compared with the Transformer, ShuffleNetV2, and YOLOv5 algorithms, the CBAM-BiLSTM can obtain more accurate power allocation coefficients and improve system performance. Compared to ShuffleNetV2, CBAM-BiLSTM reduces mean square error (MSE) by 42.8%.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0018-9545
1939-9359
DOI:10.1109/TVT.2023.3342173