Satellite-based quantum key distribution (QKD) offers a promising approach toward the realization of global-scale quantum communication networks. However, current systems are restricted to nighttime operation, as intense daytime background noise drastically degrades the signal-to-noise ratio. While single-mode fiber coupling effectively suppresses background noise, maintaining high coupling efficiency and stability amidst strong atmospheric turbulence remains a significant challenge. In this study, we propose a deep-learning-based adaptive-optics method for turbulence correction, which is experimentally validated over 1.4- and 7-km free-space channels and compared with wave-front-sensor-based systems. The proposed method achieves substantial improvements in the Strehl ratio under strong turbulence conditions. Furthermore, we numerically evaluate the performance of a free-space quantum key distribution system incorporating our deep-learning-based method, which achieved higher key rates and longer propagation distances. By directly estimating phase information from a defocused image, this approach provides an efficient and cost-effective solution for strong turbulence correction, paving the way for practical daytime QKD.

Close

PrevNext

View figure in article

• Optical quantum information processing
• Quantum cryptography
• Quantum networks
• Turbulence
• Earth's atmosphere
• Satellite communication networks
• Deep learning
• Optical techniques