Deep Learning-Based Channel Estimation for Continuous Variable Quantum Cryptography in Free-Space Optical (FSO) Communication
Gottumukkala Srisiri, Nallagorla Lokesh, Tirumala Sravan, Hima Bindu Valiveti, K. Swaraja, M Sri Uma Suseela · 2025
A network system's credibility is highly dependent on its resistance to cyber-attacks. With the advent of quantum computers, the security of traditional network systems is compromised since most of the networking systems rely on the difficulty of factoring out the prime factors which can be easily done with Shor's algorithm with the help of quantum computers. Another advantage of Quantum Key Distribution (QKD) over the traditional cryptographic system is that any attempt to eavesdrop the data introduces anomalies due to the Heisenberg uncertainty principle and the no-cloning theorem. The authors propose a Continuous Variable-QKD (CV-QKD) system where the information is encoded in the same coherent state of light that can be detected through homodyne or heterodyne detection which is compatible with standard telecom components. This CV-QKD system also has an authentication layer and a classical channel. But in this paper, we mainly focus on the quantum channel and the efficiency of using channel estimation in FSO which supposedly gives a better output values than traditional Digital Signal Processing (DSP) with equalization (Linear Interpolation). Metrics such as Mean Square Error (MSE) is used to measure the efficiency of the system.