A Hybrid Sequence Key Generation Scheme based on Channel and Pseudo-Random Information

With the growing importance of information security, numerous physical-layer encryption algorithms have been developed. However, when deployed in non-reciprocal wireless environments, these schemes often suffer from a high key error rate (KER), elevated key outage probability (KOP), and low encryption efficiency. To overcome these limitations, this paper proposes a Hybrid Sequence Key Generation (HSKG) scheme with targeted design strategies. First, the channel is mapped into a unified fixed-value subspace through channel superposition and discretization of a mapping function. The index of the upper boundary of the fixed-value interval is exchanged to alleviate channel non-reciprocity while preventing leakage of legitimate channel information. This correction significantly enhances the channel consistency, improving the correlation coefficient from 0.5623 to 0.9998. Second, an exchange factor is generated from the corrected channel sequence. The corrected sequence and a pseudo-random sequence are then block-exchanged and combined based on this factor to construct a hybrid sequence. A quadratic key extraction model is subsequently introduced to maximize the divergence between the hybrid sequence and the original sequences, thereby reducing the probability of successful eavesdropping by Eve. Finally, identity legitimacy is verified by evaluating the consistency of the decoded pseudo-random sequence across multiple stages. Extensive experimental results demonstrate the security and effectiveness of the proposed scheme. In real-world environments, the HSKG scheme achieves an outage probability (OP) of 0.0166, a KOP of 0.0486, a KER of 0.0001, and a security rate (SR) of 1.1 bit/s. Benefiting from its targeted design, the HSKG scheme significantly outperforms the BE algorithm in both connection-interruption and security-interruption performance.