Abstract
This work investigates the simultaneous transmission of a secret message (SM) and the generation of a secret key (SK) over a state-dependent wiretap channel (SD-WTC). We assume non-causal channel state information (CSI) is available at the encoder. By deriving a new inner bound on the semantic-security (SS) SM-SK capacity region, we demonstrate improvements over existing benchmarks. The proposed scheme utilizes superposition coding and a likelihood encoder, with security established through strong soft-covering lemmas. The derived region attains capacity for a certain class of SD-WTCs and outperforms previously known lower bounds even when security requirements are relaxed.
Introduction: Problem Formulation
In Physical Layer Security (PLS), the objective is to exploit inherent channel characteristics like noise and interference to ensure secrecy, moving beyond reliance on computational complexity. This paper focuses on the State-Dependent Wiretap Channel (SD-WTC), where a sender (Alice) communicates with a legitimate receiver (Bob) in the presence of an eavesdropper (Eve). The channel's behavior is influenced by a random state $S$, which is known non-causally to Alice.
Alice aims to achieve two goals simultaneously:
- Secret Message (SM) Transmission: Sending a message $M$ that Bob can decode but Eve cannot.
- Secret Key (SK) Generation: Agreeing on a key $K$ with Bob for future use, kept secret from Eve.
The fundamental challenge lies in the "Message-Key Trade-off." Channel resources must be split between coordinating with the state $S$ for reliability and masking information from Eve for security. Previous research often relied on "Weak Secrecy" metrics; this work upgrades the standard to the more rigorous Semantic Security.
Proposed Coding Scheme and Security Analysis
The authors propose a novel coding strategy based on three key pillars:
- Superposition Coding: A two-layered codebook structure is employed. The "inner layer" is designed to be decodable by the eavesdropper, effectively "saturating" her observation and leaving no capacity for her to learn about the "outer layer," which carries the sensitive SM and SK.
- Likelihood Encoder: Instead of traditional joint typicality, a likelihood-based encoding approach is used to correlate the transmitted signal $X$ with the known state $S$. This provides a more flexible and efficient way to manage the reliability-secrecy trade-off.
- Strong Soft-Covering Lemma (SCL): To ensure Semantic Security (SS)—a cryptographic gold standard that guarantees security against the worst-case message distribution—the authors use advanced versions of the SCL. This ensures that Eve's intercepted signal is statistically indistinguishable from pure noise.
Main Results and Contributions
The primary contribution is a new inner bound for the SM-SK capacity region. Key findings include:
- Superiority over Benchmarks: The derived region is strictly larger than the previous best-known result by Prabhakaran et al. (2012).
- Capacity Characterization: For specific "less-noisy" eavesdropper channels, the proposed inner bound is shown to be the exact capacity.
- Redundancy as a Resource: A significant insight is that the redundancy used for state coordination can be harvested as a secret key if Bob can decode the coordination index while Eve cannot.
- Correction of Prior Work: The authors identify and correct flaws in previous literature (e.g., Zibaeenejad 2015) regarding secret key generation rates.
Numerical Validation: Memory with Stuck-at Faults
To demonstrate the practical gains of the new bound, the authors analyze a specific channel model: a memory with stuck-at faults followed by a Binary Erasure Channel (BEC).
In this setup, the state $S$ represents the locations and values of stuck-at faults in a memory device. The encoder uses its knowledge of $S$ to write data $X$ such that the output $G$ is reliable despite the faults. The BEC then introduces further erasures before reaching the decoder. Numerical results confirm that the proposed scheme achieves higher secret rates than prior methods in this specific scenario.
Conclusion
This research advances the field of information-theoretic security by providing tighter bounds for secret message and key rates under the rigorous semantic security standard. While the inner bound is improved, finding the absolute outer bound for all state-dependent wiretap channels remains an open challenge for future study.
Summary updated