The reciprocity of wireless channels is a prerequisite for physical layer secret key generation (SKG). However, inherent factors, such as noise, asynchronous observations, and hardware impairments, disrupt the ideal reciprocity in the channel state information (CSI) observed by the two legitimate parties. To address this issue, we propose a mutual information-driven autoencoder (MIAE) architecture to extract reciprocal channel features from the non-ideal channel observations of legitimate parties. MIAE is constructed with an AutoEncoder Network (AENet) and a mutual information neural estimator (MINE). Specifically, AENet employs a structure with dual encoders and a shared decoder. The two encoders, integrated with convolutional block attention modules (CBAMs), are designed to focus on reciprocal features within CSI observations and to compensate for temporal variations caused by asynchronous observations. The shared encoder and the correspondingly designed loss function further concentrate the autoencoder’s reciprocity enhancement capability into the encoders. MINE is integrated into the proposed MIAE to estimate the mutual information between the channel features of the two parties. This estimation is then used to formulate a mutual information loss, which guides the encoders to learn channel features that closely match the optimal distribution, thereby boosting the key generation rate. Furthermore, a complete SKG scheme is designed based on the proposed channel feature extractor, MIAE. Simulation results show that our proposed MIAE can extract channel features with strong reciprocity and thus achieve excellent SKG performance based on the extracted features. The generalization performance of the proposed MIAE architecture for communication scenarios of different scales has also been examined.

Secret-key generation (SKG) at physical layer is considered as a promising solution for lightweight key distribution. On the analysis of the secret-key capacity over multipath fading channels, existing studies generally neglect the channel observation method, which has a direct impact on the secret-key capacity. Moreover, there is a lack of concise and interpretable expressions for the secret-key capacity based on the channel state information. Motivated by the above, we analyze the performance of SKG based on MMSE estimation of channel impulse response over multipath fading channels. The general expression of secret-key capacity is derived. We find that the secret-key capacity depends on the estimation signal-to-noise ratios (SNRs) of channel-tap gains. Then, the condensed-parameter expressions of exact secret-key capacity and the expressions of asymptotic secret-key capacity at high SNR are derived in two specific PDP cases. We show that the bandwidth and the transmission SNR determine the upper bound of secret-key capacity over channels with different degrees of freedom for the flat PDP case. In addition, we find that the asymptotic secret-key capacities under these two PDPs follow a uniform form, which is a linear function of the channel estimation SNR (in dB). Finally, we simulate the channel probing process under different transmission and channel parameters and exploit the simulated channel estimates to estimate the secret-key capacity through numerical methods. The simulation results demonstrate the theoretical analysis and conclusions.

Devices in the Internet of Things (IoT) are usually limited in computing resources and energy capacity, which means that encryption schemes with higher complexity are not suitable for them to ensure secure communication. As a promising solution to this problem, physical layer key generation suggests that shared secret keys can be generated from noisy wireless channel measurements to enhance the security of wireless communications. In this article, we propose a key generation scheme with extremely low implementation complexity, which allows physical layer key generation to be implemented on IoT nodes. First, we preprocess the channel measurements with simple moving average filtering before quantization to improve channel reciprocity. Next, a bidirectional difference quantization scheme is proposed to realize reliable quantization of channel measurements, which is ingenious in that the quantization process does not depend on quantization thresholds, and thus, the mismatched key bits caused by measurements close to quantization thresholds can be effectively avoided. Then, we propose an improved Cascade protocol to achieve lightweight and efficient information reconciliation. The simulation results show that our scheme can well balance the reliability and efficiency of key generation, and has excellent performance in terms of implementation complexity and key randomness.