In this paper, the authors introduce two deep joint source and channel coding (DJSCC) structures with attention modules for multi-input multi-output (MIMO) channels. The proposed structures include a serial and parallel architecture that adapt to varying channel qualities. By utilizing singular value decomposition (SVD)-based precoding, the MIMO channel is decomposed into sub-channels, allowing for improved image transmission performance. The attention modules in both architectures adjust information outputs based on channel qualities, demonstrating enhanced system adaptability. Experimental results show that the proposed DJSCC structures outperform traditional methods by transmitting more information over better sub-channels. The study also explores the impact of different sub-channel qualities on information transmission through entropy estimation.
The research focuses on developing adaptive DJSCC structures for MIMO systems using attention modules to enhance image transmission performance. By leveraging SVD-based precoding and attention mechanisms, the proposed structures demonstrate improved adaptability to varying channel qualities, resulting in enhanced data transmission efficiency.
To Another Language
from source content
arxiv.org
Key Insights Distilled From
by Weiran Jiang... at arxiv.org 03-13-2024
https://arxiv.org/pdf/2311.07041.pdfDeeper Inquiries