Shared secret key extraction from atmospheric optical wireless channels with multi-scale information reconciliation

Due to the impact of turbulence, atmospheric optical wireless channel exhibits characteristics such as time- varying, space-varying and natural randomness, which can be used as a common natural random source for shared secret key extraction. Wireless laser communication technology boasts advantages like high bandwidth and fast transmission, which is conducive to improving key generation rate. Additionally, the strong anti-interference of laser signal helps to reduce key disagreement rate. Moreover, the laser beam's good directionality effectively decreases the risk of eavesdropping on key information. Given its advantages and a scarcity of research in this regard, this paper proposes a scheme of shared secret key extraction from atmospheric optical wireless channels with multi-scale information reconciliation. In the scheme, to increase the cross-correlation coefficient of signal samples at the two legitimate parties, a preprocessing algorithm is designed based on a denoising algorithm and a threshold-based outliers elimination algorithm, and the denoising algorithm is inspired by the Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDan); moreover, a multi-level quantization algorithm based on Equilibrium-Optimizer(EO) is developed to balance and optimize distribution of sample points in the sample space; furthermore, to simplify the process of and decrease the computational complexity of information reconciliation, a concept of a multi- scale information reconciliation is formed, on the basis of which three algorithms, B-MSIR, I-MSIR and C-MSIR, are formulated. Finally, its performance is verified by numerical simulations and experiments, and the results show it has better performance in terms of the key disagreement rate, the key generation rate and the key randomness compared with several state-of-the-art algorithms.