Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

IEEE Internet of Things Journal

Secret Key Generation Assisted by Denoising Autoencoder

Zhengyu Xu, Shun Kojima, Shinya Sugiura

Published Jan 1, 2026• 0 citations
Publisher

Abstract

In this paper, we propose a denoising autoencoder (DAE)-assisted secret key generation (SKG) scheme, designed to improve the reliability and efficiency of physical-layer secret key generation. Our approach leverages the capability of DAEs to learn denoised channel representations while suppressing additive noise, thus reconstructing channel features that are more strongly correlated between the legitimate nodes. By mitigating reciprocity imperfections, the proposed scheme reduces the key disagreement rate (KDR) and improves the key generation rate (KGR). Furthermore, our extensive simulation results confirm that the DAE-assisted SKG outperforms the transform-based and learning-based baselines, including discrete wavelet transform, principal component analysis, discrete cosine transform, traditional autoencoder (AE)-based approaches, convolutional neural network (CNN)-based approaches, convergence-learning fully connected neural network, and mutual-information-driven AE. The NIST SP 800-22 results confirm that the generated keys successfully pass the suite of statistical randomness tests after privacy amplification. These results indicate that the proposed DAE-assisted framework is an effective preprocessing approach for empirical SKG under the adopted threat model and assumptions, with moderate online complexity relative to the learning-based baselines.

DOI
10.1109/jiot.2026.3739810

Related papers

Denoising-Autoencoder-Assisted Physical Layer Secret Key Generation

Zhengyu Xu, Shun Kojima, Shinya Sugiura

In this paper, we propose denoising autoencoder (DAE)-assisted secret key generation (SKG), where channel noise reciprocity imperfections induced due to wireless channel measurements are suppressed, hence significantly enhancing the reliability and efficiency. More specifically, the DAE is capable of capturing the intrinsic structure of input distributions, reconstructing the original data structure, and removing additive noise while preserving the essential structure of signals. In our analysis, it is demonstrated that the proposed SKG scheme exhibits higher performance than the conventional schemes in terms of key disagreement rate (KDR), secret key capacity (SKC), and randomness of the generated keys.

arXiv:2607.14505•International Conference on Communications•May 24, 2026•0 citations

Autoencoder-Based Noise Augmentation for Physical Layer Secret Key Generation in IoT Networks

Anas Alashqar, Ehsan Olyaei Torshizi, Raed Mesleh +1 more

Physical Layer Secret Key Generation (PSKG) exploits wireless channel reciprocity to establish cryptographic keys between legitimate users, providing a promising solution for securing IoT networks. However, reciprocity in time-division duplex (TDD) systems is often compromised by hardware imperfections and channel noise, posing significant challenges for reliable key generation. To address these issues, this paper proposes an autoencoder-based noise augmentation (AE-NA) model that operates in an unsupervised manner. Specifically, the proposed model integrates dynamic Gaussian noise augmentation directly into the autoencoder architecture. This approach enhances feature extraction capabilities and improves noise robustness, enabling the network to learn noise-invariant representations effectively from noisy channel estimates. Furthermore, we present a comprehensive PSKG framework that incorporates the proposed AE-NA model. Reported results demonstrate that the AE-NA-based PSKG significantly reduces the key disagreement ratio (KDR) while enhancing key randomness, thereby providing a scalable and effective solution for secure key generation in dynamic IoT environments.

IEEE International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC)•Sep 1, 2025•0 citations

Sparse Autoencoder-Driven Key Generation for Enhanced IoT Security

Anas Alashqar, Raed Mesleh, Ehsan Olyaei Torshizi +1 more

Physical-layer secret key generation (PSKG) is a well-known and effective method for boosting wireless security in the Internet of Things (IoT). This technique creates cryptographic keys from correlated wireless channel measurements by taking advantage of the reciprocal nature of uplink and downlink channels. However, in time-division duplex (TDD) systems, issues such as hardware imperfection and noise can disrupt channel reciprocity, making the key generation process more difficult. This paper presents a new PSKG approach that tackles these issues by using deep learning to enhance channel reciprocity, even with imperfect channel state information (CSI). Specifically, a denoising sparse autoencoder (DSAE) is introduced to efficiently learn and extract channel reciprocity features in TDD OFDM systems. Furthermore, this study presents a low-complexity DSAE-based PSKG scheme that utilizes the phase of channel frequency responses (CFR) for key generation. The reported results indicate that the proposed DSAE-based PSKG method notably decreases the key disagreement ratio (KDR) while improving randomness, offering a robust and practical solution for securing wireless communications in IoT environments.

IEEE Globecom Workshops (GC Wkshps)•Dec 8, 2024•4 citations

Physical Layer Secret Key Generation Based on Mutual Information-Driven Autoencoder

Dengke Guo, Jun Xiong, Dongtang Ma +2 more

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.

IEEE Transactions on Wireless Communications•May 1, 2025•2 citations

Physical Layer Secret Key Generation Based on Autoencoder for Weakly Correlated Channels

Jingyuan Han, Xin Zeng, Xiaoping Xue +1 more

The reciprocity of the wireless channel is the basis of the secret key generation on the physical layer as an emerging method for information security enhancement. Due to the duplex mode of communication systems, the channel responses detected by two connected legitimate nodes are not completely reciprocal. In this paper, an efficient physical layer secret key generation scheme is proposed based on an autoencoder to extract reciprocal features from weakly correlated channel estimates. The encoder part after training is equipped at both nodes to generate features for quantization in various channel environments. Simulation results show lower mean squared error (MSE) and better mutual information of the extracted features comparing to the most efficient linear generation method, the principal component analysis (PCA) based one. It reveals the great potential to generate symmetric secret keys with better performance, particularly in low SNR scenarios.

IEEE/CIC International Conference on Communications in China (ICCC)•Aug 9, 2020•28 citations

Performance Improvement of Secret Key Generation Scheme in Wireless Indoor Environment

Mike Yuliana, Wirawan, Suwadi

The Secret Key Generation (SKG) scheme that exploits the reciprocity and uniqueness of wireless channel between two users plays a significant part in a new increasing distributed security system. The scheme performance can be distinguished based on the low value of Key disagreement Rate (KDR), the high value of Key Generation Rate (KGR), as well as the fulfillment of the NIST randomness standard. The previous SKG scheme has a high KDR due to a direct quantization of a measurement result of the Received Signal Strength (RSS). To overcome the above issue, we conduct a pre-processing of measurement result before quantization with the Kalman method. The pre-process is carried out to improve the channel reciprocity between two legitimate users with the objective to reduce the bit mismatch. Through an experiment, we propose a new quantization scheme called a Modified Multi-Bit (MMB) that uses a multi-bit system on every level of quantization. The test results show that the proposed combination of preprocessing and the MMB scheme has a better performance compared to the existing schemes in terms of KDR and KGR. The Secret Key generated by our scheme also fulfills the NIST randomness standard.

International Journal of Communication Networks and Information Security (IJCNIS)•Apr 17, 2022•20 citations
In this paper, we propose a denoising autoencoder (DAE)-assisted secret key generation (SKG) scheme, designed to improve the reliability and efficiency of physical-layer secret key generation. Our approach leverages the capability of DAEs to learn denoised channel representations while suppressing additive noise, thus reconstructing channel features that are more strongly correlated between the legitimate nodes. By mitigating reciprocity imperfections, the proposed scheme reduces the key disagreement rate (KDR) and improves the key generation rate (KGR). Furthermore, our extensive simulation results confirm that the DAE-assisted SKG outperforms the transform-based and learning-based baselines, including discrete wavelet transform, principal component analysis, discrete cosine transform, traditional autoencoder (AE)-based approaches, convolutional neural network (CNN)-based approaches, convergence-learning fully connected neural network, and mutual-information-driven AE. The NIST SP 800-22 results confirm that the generated keys successfully pass the suite of statistical randomness tests after privacy amplification. These results indicate that the proposed DAE-assisted framework is an effective preprocessing approach for empirical SKG under the adopted threat model and assumptions, with moderate online complexity relative to the learning-based baselines.
In this paper, we propose denoising autoencoder (DAE)-assisted secret key generation (SKG), where channel noise reciprocity imperfections induced due to wireless channel measurements are suppressed, hence significantly enhancing the reliability and efficiency. More specifically, the DAE is capable of capturing the intrinsic structure of input distributions, reconstructing the original data structure, and removing additive noise while preserving the essential structure of signals. In our analysis, it is demonstrated that the proposed SKG scheme exhibits higher performance than the conventional schemes in terms of key disagreement rate (KDR), secret key capacity (SKC), and randomness of the generated keys.
Physical Layer Secret Key Generation (PSKG) exploits wireless channel reciprocity to establish cryptographic keys between legitimate users, providing a promising solution for securing IoT networks. However, reciprocity in time-division duplex (TDD) systems is often compromised by hardware imperfections and channel noise, posing significant challenges for reliable key generation. To address these issues, this paper proposes an autoencoder-based noise augmentation (AE-NA) model that operates in an unsupervised manner. Specifically, the proposed model integrates dynamic Gaussian noise augmentation directly into the autoencoder architecture. This approach enhances feature extraction capabilities and improves noise robustness, enabling the network to learn noise-invariant representations effectively from noisy channel estimates. Furthermore, we present a comprehensive PSKG framework that incorporates the proposed AE-NA model. Reported results demonstrate that the AE-NA-based PSKG significantly reduces the key disagreement ratio (KDR) while enhancing key randomness, thereby providing a scalable and effective solution for secure key generation in dynamic IoT environments.
Physical-layer secret key generation (PSKG) is a well-known and effective method for boosting wireless security in the Internet of Things (IoT). This technique creates cryptographic keys from correlated wireless channel measurements by taking advantage of the reciprocal nature of uplink and downlink channels. However, in time-division duplex (TDD) systems, issues such as hardware imperfection and noise can disrupt channel reciprocity, making the key generation process more difficult. This paper presents a new PSKG approach that tackles these issues by using deep learning to enhance channel reciprocity, even with imperfect channel state information (CSI). Specifically, a denoising sparse autoencoder (DSAE) is introduced to efficiently learn and extract channel reciprocity features in TDD OFDM systems. Furthermore, this study presents a low-complexity DSAE-based PSKG scheme that utilizes the phase of channel frequency responses (CFR) for key generation. The reported results indicate that the proposed DSAE-based PSKG method notably decreases the key disagreement ratio (KDR) while improving randomness, offering a robust and practical solution for securing wireless communications in IoT environments.
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.
The reciprocity of the wireless channel is the basis of the secret key generation on the physical layer as an emerging method for information security enhancement. Due to the duplex mode of communication systems, the channel responses detected by two connected legitimate nodes are not completely reciprocal. In this paper, an efficient physical layer secret key generation scheme is proposed based on an autoencoder to extract reciprocal features from weakly correlated channel estimates. The encoder part after training is equipped at both nodes to generate features for quantization in various channel environments. Simulation results show lower mean squared error (MSE) and better mutual information of the extracted features comparing to the most efficient linear generation method, the principal component analysis (PCA) based one. It reveals the great potential to generate symmetric secret keys with better performance, particularly in low SNR scenarios.
The Secret Key Generation (SKG) scheme that exploits the reciprocity and uniqueness of wireless channel between two users plays a significant part in a new increasing distributed security system. The scheme performance can be distinguished based on the low value of Key disagreement Rate (KDR), the high value of Key Generation Rate (KGR), as well as the fulfillment of the NIST randomness standard. The previous SKG scheme has a high KDR due to a direct quantization of a measurement result of the Received Signal Strength (RSS). To overcome the above issue, we conduct a pre-processing of measurement result before quantization with the Kalman method. The pre-process is carried out to improve the channel reciprocity between two legitimate users with the objective to reduce the bit mismatch. Through an experiment, we propose a new quantization scheme called a Modified Multi-Bit (MMB) that uses a multi-bit system on every level of quantization. The test results show that the proposed combination of preprocessing and the MMB scheme has a better performance compared to the existing schemes in terms of KDR and KGR. The Secret Key generated by our scheme also fulfills the NIST randomness standard.