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

International Conference on Future and Intelligent Networking (FINE)

Towards Eavesdropper Awareness in CSI-Based Key Generation

Ashikur Rahman Nobel, Honggang Wang, Hua Fang

Published Aug 1, 2026• 0 citations
Publisher

Abstract

Physical-layer key generation derives shared bit strings from reciprocal wireless measurements, but conventional feature selection often optimizes Alice-Bob agreement without explicitly assessing whether the selected features are also observable by a passive eavesdropper. This paper presents a compact channel-state-information (CSI) key-generation pipeline that selects discrete-cosine-transform (DCT) coordinates using an eavesdropper-aware score and compares it with a capacity-matched eavesdropper-unaware control. Both variants share the same 96-dimensional CSI feature contract, 36-coordinate budget, guarded 2-bit Gray quantization, reconciliation, confirmation, Toeplitz hashing, grouped split, and development-derived extraction accounting. The study uses simultaneous recordings from two passive eavesdroppers over static, walking, endpoint-motion, line-of-sight (LOS), non-line-of-sight (NLOS), and placement conditions. Across 50 held-out capture sets, the eavesdropper-aware selector yields 26,971 final bits at 35.961 bit/s and a maximum post-generation Eve correlation of 0.003956; the matched control yields 27,873 bits at 37.164 bit/s and 0.005341. The results show a measured rate-exposure trade-off: Eve-aware selection lowers the evaluated exposure indicator while retaining the same number of confirmed sessions.

DOI
10.1109/fine71157.2026.00035

Related papers

Let’s shake on it: Extracting secure shared keys from Wi-Fi CSI

Tomer Avrahami, Ofer Amrani, Avishai Wool

A shared secret key is necessary for encrypted communications. Since Wi-Fi relies on OFDM, we suggest a method for generating such key by utilizing Wi-Fi’s channel state information (CSI). The channel is typically reciprocal but very sensitive to location: while (legitimate) Alice and Bob should observe the same CSI, an eavesdropper Eve shall observe uncorrelated CSI when positioned over 0.5 wavelength away from either node.We show that if endpoint Bob is shaken, sufficient diversity is induced in the CSI, thus rendering it a true source for randomness. We then show that the CSI among neighboring sub-carriers is correlated, so we select a small set of judiciously-spaced sub-carriers, and decide on a value based on a majority rule about each. Consequently, Alice and Bob observe a 5-15% bit mismatch rate (BMR) in the extracted bitstream while Eve experiences a BMR of around 50% even when placed 10cm away from Alice.We employ the cryptography-oriented definition of min-entropy to estimate the number of secure bits in the received bitstream, and employ the so-called Cascade algorithm of quantum-key-distribution to reconcile Alice and Bob’s bitstreams while quantifying the number of bits leaked by the algorithm. Accounting for both the min-entropy and the Cascade-based leakage we quantify the Secured Bit Generation Rate realized by our method.Extensive set of experiments have been conducted in an indoor environment revealing secure bit generation rate of 1.2–1.6 bits per packet, for endpoint distances ranging from 0.5m–9m; a (128-bit) secured shared-key is generated in 20sec of device shaking.

arXiv:2307.05423•IEEE International Conference on Mobile Ad-Hoc and Sensor Systems (MASS)•Sep 25, 2023•2 citations

Physical Layer Security against an Informed Eavesdropper in Underwater Acoustic Channels: Feature Extraction and Quantization

Konstantinos Pelekanakis, Seckin Anil Yildirim, Georgios Sklivanitis +3 more

During the Rapid Environmental Picture 2018 (REP18) sea trial, two underwater acoustic nodes (Alice and Bob) exchanged 897 channel probes over different ranges and environmental conditions. In this short paper, Alice and Bob independently process their received probes offline with the aim to generate a cryptographic key based on Physical Layer Security (PLS). Using their estimated Channel Impulse Responses (CIRs), they compute and quantize four pre-agreed channel features. Eve is a simulated eavesdropper who is aware of the PLS algorithm, the 3D positions of Alice and Bob and the acoustic properties of the environment. Eve uses the de facto standard Bellhop acoustic simulator to predict the bi-directional CIRs between Alice and Bob and compute her own quantized features. We calculate the Bit Disagreement Ratio (BDR), which is a function of the number of disagreeing bits between a pair of nodes. Our results confirm that the proposed features are robust enough to yield a lower BDR between Alice and Bob than that for Eve. The BDR impact on reconciliation and secret key generation is studied in a subsequent paper [1].

Underwater Communications and Networking Conference (UComms)•Aug 31, 2021•31 citations

Physical Layer Security against an Informed Eavesdropper in Underwater Acoustic Channels: Reconciliation and Privacy Amplification

George Sklivanitis, Konstantinos Pelekanakis, Seckin Anil Yildirim +3 more

We propose a strategy for generating the same crypto-key between two trusted underwater acoustic nodes (Alice and Bob) without revealing it to an eavesdropper (Eve). Our work builds upon the results of [1] where a methodology for generating a string of bits for Alice, Bob and Eve based on channel feature extraction and quantization is discussed. In this paper, we aim to reconcile the respective bits of Alice and Bob while minimizing the information leaked to Eve. To this end, we examine various Reed Solomon (RS) codes and measure the reconciliation rate of Alice, Bob and Eve. Additionally, we propose the Secure Hash Algorithm-3 (SHA-3) as means to eliminate any information that Eve acquires during reconciliation. We evaluate our reconciliation and privacy amplification strategies with bits generated from real underwater acoustic channel probe exchanges between Alice and Bob and Bellhop-simulated channels for Eve. Our analysis confirms that appropriate combinations of channel features and RS codes lead to a computationally secure generation of a 256-bit crypto-key according to the principles of the National Institute of Standards and Technology (NIST), even if Eve is informed about the RS encoder and the SHA-3 function.

Underwater Communications and Networking Conference (UComms)•Aug 31, 2021•32 citations

Measured statistics of reciprocal channel key generation of indoor MIMO channels

R. K. Sharma, J W Wallace

There is growing interest in physical-layer security methods that exploit the random nature of the physical propagation channel to strengthen existing crypto-systems. In reciprocal channel key generation (RCKG), legitimate nodes (Alice and Bob) observe a common fluctuating channel to generate keys that are safe from the eavesdropper (Eve). Previous work [1, 2] derived expressions for the available key bits and those safe from an eavesdropper for MIMO systems with correlated complex Gaussian statistics. This paper applies these expressions to new MIMO measurements in indoor LOS and NLOS environments, indicating how propagation effects limit secure key generation and what key generation rates can be expected in practice.

Digest - IEEE Antennas and Propagation Society. International Symposium•Jul 1, 2010•1 citation

Experimental Investigation of MIMO Reciprocal Channel Key Generation

Jon W. Wallace, R. K. Sharma

There is growing interest in physical-layer key generation schemes that provide very strong or even perfect security in wireless communication systems. One such scheme is reciprocal channel key generation, where two nodes quantize reciprocal channel state information to generate keys. Although the use of multiple-input multiple-output (MIMO) techniques is interesting in this case since the number of random parameters available for rapid key generation is increased, MIMO techniques may also provide more information to an eavesdropper. This work presents a new MIMO measurement campaign performed in LOS and NLOS indoor environments that studies the correlation of the channel between legitimate users with the eavesdropper channel, revealing what key generation rates can be attained in practice and what fraction of generated key bits are safe from eavesdroppers. The effect of eavesdropper separation, number of antennas, eavesdropper advantage, and covariance separability are studied.

International Conference on Communications•May 1, 2010•10 citations

Physical Layer Secret Key Generation in Static Environments

Nasser Aldaghri, Hessam Mahdavifar

Two legitimate parties, referred to as Alice and Bob, wish to generate secret keys from the wireless channel in the presence of an eavesdropper, referred to as Eve, in order to use such keys for encryption and decryption. In general, the secret key rate highly depends on the coherence time of the channel. In particular, a straightforward method of generating secret keys in static environments results in ultra-low rates. In order to resolve this problem, we introduce a low-complexity method called induced randomness. In this method, Alice and Bob independently generate local randomness to be used together with the uniqueness of the wireless channel coefficients in order to enable high-rate secret key generation. In this work, two scenarios are considered: first, when Alice and Bob share a direct communication channel, and second, when Alice and Bob do not have a direct link and communicate through an untrusted relay. After exchanging the induced randomness, post-processing is done by Alice and Bob to generate highly-correlated samples that are used for the key generation. Such samples are then converted into bits, disparities between the sequences generated by Alice and Bob are mitigated, and the resulting sequences are then hashed to compensate for the information leakage to the eavesdropper and to allow consistency checking of the generated key bit sequences. We utilize semantic security measures and information-theoretic inequalities to upper bound the probability of successful eavesdropping attack in terms of the mutual information measures that can be numerically computed. Given certain reasonable system parameters this bound is numerically evaluated to be 2-31and 2-10.57in the first and the second scenario, respectively.

arXiv:1908.03637•IEEE Transactions on Information Forensics and Security•Jan 1, 2020•202 citations
Physical-layer key generation derives shared bit strings from reciprocal wireless measurements, but conventional feature selection often optimizes Alice-Bob agreement without explicitly assessing whether the selected features are also observable by a passive eavesdropper. This paper presents a compact channel-state-information (CSI) key-generation pipeline that selects discrete-cosine-transform (DCT) coordinates using an eavesdropper-aware score and compares it with a capacity-matched eavesdropper-unaware control. Both variants share the same 96-dimensional CSI feature contract, 36-coordinate budget, guarded 2-bit Gray quantization, reconciliation, confirmation, Toeplitz hashing, grouped split, and development-derived extraction accounting. The study uses simultaneous recordings from two passive eavesdroppers over static, walking, endpoint-motion, line-of-sight (LOS), non-line-of-sight (NLOS), and placement conditions. Across 50 held-out capture sets, the eavesdropper-aware selector yields 26,971 final bits at 35.961 bit/s and a maximum post-generation Eve correlation of 0.003956; the matched control yields 27,873 bits at 37.164 bit/s and 0.005341. The results show a measured rate-exposure trade-off: Eve-aware selection lowers the evaluated exposure indicator while retaining the same number of confirmed sessions.
A shared secret key is necessary for encrypted communications. Since Wi-Fi relies on OFDM, we suggest a method for generating such key by utilizing Wi-Fi’s channel state information (CSI). The channel is typically reciprocal but very sensitive to location: while (legitimate) Alice and Bob should observe the same CSI, an eavesdropper Eve shall observe uncorrelated CSI when positioned over 0.5 wavelength away from either node.We show that if endpoint Bob is shaken, sufficient diversity is induced in the CSI, thus rendering it a true source for randomness. We then show that the CSI among neighboring sub-carriers is correlated, so we select a small set of judiciously-spaced sub-carriers, and decide on a value based on a majority rule about each. Consequently, Alice and Bob observe a 5-15% bit mismatch rate (BMR) in the extracted bitstream while Eve experiences a BMR of around 50% even when placed 10cm away from Alice.We employ the cryptography-oriented definition of min-entropy to estimate the number of secure bits in the received bitstream, and employ the so-called Cascade algorithm of quantum-key-distribution to reconcile Alice and Bob’s bitstreams while quantifying the number of bits leaked by the algorithm. Accounting for both the min-entropy and the Cascade-based leakage we quantify the Secured Bit Generation Rate realized by our method.Extensive set of experiments have been conducted in an indoor environment revealing secure bit generation rate of 1.2–1.6 bits per packet, for endpoint distances ranging from 0.5m–9m; a (128-bit) secured shared-key is generated in 20sec of device shaking.
During the Rapid Environmental Picture 2018 (REP18) sea trial, two underwater acoustic nodes (Alice and Bob) exchanged 897 channel probes over different ranges and environmental conditions. In this short paper, Alice and Bob independently process their received probes offline with the aim to generate a cryptographic key based on Physical Layer Security (PLS). Using their estimated Channel Impulse Responses (CIRs), they compute and quantize four pre-agreed channel features. Eve is a simulated eavesdropper who is aware of the PLS algorithm, the 3D positions of Alice and Bob and the acoustic properties of the environment. Eve uses the de facto standard Bellhop acoustic simulator to predict the bi-directional CIRs between Alice and Bob and compute her own quantized features. We calculate the Bit Disagreement Ratio (BDR), which is a function of the number of disagreeing bits between a pair of nodes. Our results confirm that the proposed features are robust enough to yield a lower BDR between Alice and Bob than that for Eve. The BDR impact on reconciliation and secret key generation is studied in a subsequent paper [1].
We propose a strategy for generating the same crypto-key between two trusted underwater acoustic nodes (Alice and Bob) without revealing it to an eavesdropper (Eve). Our work builds upon the results of [1] where a methodology for generating a string of bits for Alice, Bob and Eve based on channel feature extraction and quantization is discussed. In this paper, we aim to reconcile the respective bits of Alice and Bob while minimizing the information leaked to Eve. To this end, we examine various Reed Solomon (RS) codes and measure the reconciliation rate of Alice, Bob and Eve. Additionally, we propose the Secure Hash Algorithm-3 (SHA-3) as means to eliminate any information that Eve acquires during reconciliation. We evaluate our reconciliation and privacy amplification strategies with bits generated from real underwater acoustic channel probe exchanges between Alice and Bob and Bellhop-simulated channels for Eve. Our analysis confirms that appropriate combinations of channel features and RS codes lead to a computationally secure generation of a 256-bit crypto-key according to the principles of the National Institute of Standards and Technology (NIST), even if Eve is informed about the RS encoder and the SHA-3 function.
There is growing interest in physical-layer security methods that exploit the random nature of the physical propagation channel to strengthen existing crypto-systems. In reciprocal channel key generation (RCKG), legitimate nodes (Alice and Bob) observe a common fluctuating channel to generate keys that are safe from the eavesdropper (Eve). Previous work [1, 2] derived expressions for the available key bits and those safe from an eavesdropper for MIMO systems with correlated complex Gaussian statistics. This paper applies these expressions to new MIMO measurements in indoor LOS and NLOS environments, indicating how propagation effects limit secure key generation and what key generation rates can be expected in practice.
There is growing interest in physical-layer key generation schemes that provide very strong or even perfect security in wireless communication systems. One such scheme is reciprocal channel key generation, where two nodes quantize reciprocal channel state information to generate keys. Although the use of multiple-input multiple-output (MIMO) techniques is interesting in this case since the number of random parameters available for rapid key generation is increased, MIMO techniques may also provide more information to an eavesdropper. This work presents a new MIMO measurement campaign performed in LOS and NLOS indoor environments that studies the correlation of the channel between legitimate users with the eavesdropper channel, revealing what key generation rates can be attained in practice and what fraction of generated key bits are safe from eavesdroppers. The effect of eavesdropper separation, number of antennas, eavesdropper advantage, and covariance separability are studied.
Two legitimate parties, referred to as Alice and Bob, wish to generate secret keys from the wireless channel in the presence of an eavesdropper, referred to as Eve, in order to use such keys for encryption and decryption. In general, the secret key rate highly depends on the coherence time of the channel. In particular, a straightforward method of generating secret keys in static environments results in ultra-low rates. In order to resolve this problem, we introduce a low-complexity method called induced randomness. In this method, Alice and Bob independently generate local randomness to be used together with the uniqueness of the wireless channel coefficients in order to enable high-rate secret key generation. In this work, two scenarios are considered: first, when Alice and Bob share a direct communication channel, and second, when Alice and Bob do not have a direct link and communicate through an untrusted relay. After exchanging the induced randomness, post-processing is done by Alice and Bob to generate highly-correlated samples that are used for the key generation. Such samples are then converted into bits, disparities between the sequences generated by Alice and Bob are mitigated, and the resulting sequences are then hashed to compensate for the information leakage to the eavesdropper and to allow consistency checking of the generated key bit sequences. We utilize semantic security measures and information-theoretic inequalities to upper bound the probability of successful eavesdropping attack in terms of the mutual information measures that can be numerically computed. Given certain reasonable system parameters this bound is numerically evaluated to be 2-31and 2-10.57in the first and the second scenario, respectively.