本文针对高斯窃听信道模型下的物理层安全问题,提出一种基于删余Polar码的、拥有编码比特的可信度计算的物理层安全编码方案。方案考虑到信道噪声对译码结果的影响,通过Bhattacharyya参数评估各信息比特的恢复差错概率,并将这恢复差错概率应用于编码比特的可信度计算。理论证明所提出的方案可使窃听者保持较高误码率,同时合法用户在高于自身信噪比门限时保证较低误码率。数值仿真结果表明所提出的安全编码方案能够有效地减小安全间隙。当窃听信道质量比合法信道质量稍差时窃听者误码率能迅速逼近0.5,而合法信道的误码率能够降到10-5以下,大大地降低了Bob和Eve间的安全间隙。
A physical layer secrecy coding scheme based on punctured polar codes is proposed in the paper,which aims at the physical layer secure problem under Gaussian wiretap channel,where the punctured position is calculated by the reliability of code-bit. Considering different effects of channel noise on the decoding results,the recovery-error probability of information-bits related to the Bhattacharyya parameter of Gaussian channel is useful for calculating the reliability of code-bit,which contributes to the decision of the punctured positions. Theoretical analysis shows that the eavesdropper has high bit error rate(BER) under SNR threshold,while the authorized user can keep low BER when the noise is over the SNR threshold. Numerical simulation results show that the proposed puncturing scheme reduces the security gap efficiently. The BER of eavesdropper gets close to 0. 5 rapidly even though his channel is only a little worse than the authorized channel,while the BER for the authorized channel could fall below a 10- 5. The gap for security between Bob and Eve reduces greatly.