UAV-Deployed OAM-BB84 QKD: Turbulence- and Misalignment-Resilient Decoy-State Finite-Key Security with AI-Assisted Calibration

TL;DR

This work addresses secure quantum key distribution over mobile UAV links using OAM-encoded BB84, integrating a unified turbulence/misalignment channel model with decoy-state finite-key analysis and a lightweight physics-informed AI calibration to stabilize performance. The approach couples AO-enabled channel stabilization, decoy-state security with composable finite-key bounds, and an AI module that classifies pulses and guides decoding, yielding robust key rates under nonstationary flight conditions. Key contributions include a complete evaluation framework (system diagram, turbulence/QBER maps, decoy/finite-key analyses, and AI calibration metrics) and quantitative evidence that AI-assisted calibration can enhance secret-key rates by 10–30% in moderate turbulence and jitter scenarios while preserving security. The practical impact lies in providing a clear pathway for airborne OAM–QKD deployment, outlining operating envelopes, design guidelines, and field-trial roadmap to realize composably secure, mobile quantum networks.

Abstract

We present a theoretical framework for quantum key distribution (QKD) using orbital angular momentum (OAM) encoded BB84 on an unmanned aerial vehicle (UAV) platform. A unified channel model captures Kolmogorov turbulence, pointing induced misalignment, and finite aperture clipping, enabling quantitative predictions of inter mode crosstalk and the resulting quantum bit error rate (QBER). Using a weak plus vacuum decoy state formulation, we derive composable finite key lower bounds on the secret key rate that incorporate statistical fluctuations, detector dark counts, efficiency mismatch, and error correction leakage. To stabilize performance under non stationary flight conditions, we introduce a lightweight physics informed learning module that combines physical priors with measured link statistics to classify valid pulses, reject corrupted data, and recommend decoding strategies. We outline a complete evaluation pipeline including UAV system architecture, turbulence driven QBER maps, decoy optimization, finite key scaling, and AI calibration metrics. Simulations indicate that under moderate turbulence and milliradian level pointing jitter, the proposed AI assisted method can improve the secret key rate by 10 percent to 30 percent while preserving composable security.

Figures (7)

• Figure 1: UAV-deployable OAM–BB84 system. transmitter (UAV-A) with decoy modulation, OAM encoder, and PAT. Middle: free-space channel with turbulence ($r_0$), pointing jitter ($\sigma_\theta$), aperture clipping, and induced inter-mode crosstalk $P(\ell'|\ell)$. receiver (UAV-B/ground) with AO, mode sorter, detectors, AI-assisted calibration, and composable finite-key post-processing over an authenticated classical channel.
• Figure 2: QBER vs. Fried parameter $r_0$ and pointing jitter $\sigma_\theta$. Smaller $r_0$ (stronger turbulence) and larger $\sigma_\theta$ drive inter-mode crosstalk and elevate the error floor.
• Figure 3: Decoy-state OAM--BB84: secret key rate vs. distance for three $(\mu_s,\mu_w)$ pairs (weak+vacuum). Curves illustrate the trade-off between higher detection probability (larger $\mu_s$) and vulnerability to background/dark-count-limited error floors.
• Figure 4: Finite-key effect at 8 km: secret key rate $R(n)$ vs. block length. The dashed line indicates the asymptotic rate $R_\infty$; finite-size penalties dominate for small $n$ and vanish as $n$ increases.
• Figure 5: ROC comparison of the physics-informed AI classifiers. Gradient Boosting achieves the best trade-off (AUC $=0.983$).