Channel-Blind Joint Source–Channel Coding for Wireless Image Transmission
Sensors · 20.06.2024 · 24 (12) · 4005
DOI: 10.3390/s24124005 ↗Kurz gefasst
Fachartikel in der Zeitschrift Sensors, online erschienen am 20.06.2024. Verfasst von Hongjie Yuan und 3 weiteren Autor:innen. Beteiligte Einrichtung: State Key Laboratory of Media Convergence & Communication, Communication University of China, Beijing 100024, China. Die Arbeit wurde bisher 26-mal zitiert (Crossref) und nennt 61 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Fundamental Research Funds.
Abstract (Originalsprache)
Joint source–channel coding (JSCC) based on deep learning has shown significant advancements in image transmission tasks. However, previous channel-adaptive JSCC methods often rely on the signal-to-noise ratio (SNR) of the current channel for encoding, which overlooks the neural network’s self-adaptive capability across varying SNRs. This paper investigates the self-adaptive capability of deep learning-based JSCC models to dynamically changing channels and introduces a novel method named Channel-Blind JSCC (CBJSCC). CBJSCC leverages the intrinsic learning capability of neural networks to self-adapt to dynamic channels and diverse SNRs without relying on external SNR information. This approach is advantageous, as it is not affected by channel estimation errors and can be applied to one-to-many wireless communication scenarios. To enhance the performance of JSCC tasks, the CBJSCC model employs a specially designed encoder–decoder. Experimental results show that CBJSCC outperforms existing channel-adaptive JSCC methods that depend on SNR estimation and feedback, both in additive white Gaussian noise environments and under slow Rayleigh fading channel conditions. Through a comprehensive analysis of the model’s performance, we further validate the robustness and adaptability of this strategy across different application scenarios, with the experimental results providing strong evidence to support this claim.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- State Key Laboratory of Media Convergence & Communication, Communication University of China, Beijing 100024, China.
Förderung
- Fundamental Research Funds
Zitieren
APA
Yuan, H., Xu, W., Wang, Y., & Wang, X. (2024). Channel-Blind Joint Source–Channel Coding for Wireless Image Transmission. Sensors, 24(12), 4005. https://doi.org/10.3390/s24124005
Vancouver
Yuan H, Xu W, Wang Y, Wang X. Channel-Blind Joint Source–Channel Coding for Wireless Image Transmission. Sensors. 2024;24(12):4005. doi:10.3390/s24124005
BibTeX
@article{Yuan2024,
author = {Yuan, Hongjie and Xu, Weizhang and Wang, Yuhuan and Wang, Xingxing},
title = {Channel-Blind Joint Source–Channel Coding for Wireless Image Transmission},
journal = {Sensors},
year = {2024},
volume = {24},
number = {12},
pages = {4005},
doi = {10.3390/s24124005}
}
EUVIMED