Distribution-Preserving Integrated Sensing and Communication with Secure Reconstruction

Yiqi Chen, Tobias J. Oechtering, Holger Boche, Mikael Skoglund, Yuan Luo

Abstract

Integrated Sensing and Communication (ISAC) is a cornerstone of 6G architecture, yet it faces significant challenges regarding statistical fidelity and data privacy. This paper investigates a framework for distribution-preserving ISAC that ensures reconstructed environmental data maintains its original statistical properties while remaining secure from eavesdroppers. By leveraging information-theoretic tools such as Common Randomness (CR) and Total Variation distance, the authors establish fundamental limits and achievable regions for communication rate, sensing distortion, and security.

1. Introduction

The evolution toward 6G networks necessitates the convergence of communication and sensing into a unified framework known as Integrated Sensing and Communication (ISAC). In this paradigm, the same spectral resources and hardware are utilized for both data transmission and environmental mapping (e.g., radar-like sensing). While traditional research has focused on minimizing distortion—typically measured by Mean Squared Error (MSE)—this approach often overlooks the statistical integrity of the sensed data and the inherent security risks of broadcasting sensing-capable signals.

This research addresses two critical gaps: the risk of distribution mismatch in downstream environment-aware algorithms and the vulnerability of environmental data to unauthorized reconstruction by eavesdroppers.

2. The Distribution-Preservation Paradigm

A primary contribution of this work is the shift from simple distortion metrics to distribution-preserving reconstruction. In many 6G applications, sensed data serves as an input for secondary algorithms, such as beamforming optimization or autonomous navigation. If the reconstruction process alters the statistical distribution of the sensed environment, these downstream models—often optimized for specific statistical properties—may perform suboptimally or fail entirely.

To mitigate this, the authors utilize Total Variation (TV) distance as a constraint. By ensuring the reconstructed sequence's distribution is nearly identical to the original state distribution, the system maintains the "statistical truth" of the environment, ensuring compatibility with subsequent processing stages.

3. Securing the Sensing Channel: The Henchman Model

Because ISAC signals are inherently broadcast, an eavesdropper within range can potentially intercept the signal to perform their own environmental sensing. This poses a significant privacy threat. To model and counter this, the researchers introduce the "Henchman and Eavesdropper" model.

In this scenario, an adversary (the Eavesdropper) attempts to reconstruct the environmental state, potentially aided by a "Henchman" who provides a limited-rate description of the data. The goal of the proposed framework is to define a "rate-distortion" region where the legitimate receiver achieves a high-fidelity, distribution-preserved estimate, while the eavesdropper is forced into a high-distortion regime, rendering their reconstructed data useless.

4. Information-Theoretic Methodology

The study employs several sophisticated information-theoretic tools to derive fundamental limits:

  • Common Randomness (CR)-Assisted Coding: The authors demonstrate that shared secret bits between the encoder and decoder can be used to "shape" the reconstructed signal. This allows for distribution preservation without requiring additional communication bandwidth.
  • Likelihood Encoding: By using likelihood encoders and the "soft-covering lemma," the authors prove the achievability of specific communication rates and sensing accuracies.
  • Stochastic vs. Deterministic Encoders: The research clarifies the role of private randomness. While stochastic encoders are beneficial, the paper proves that in deterministic channels, the communication rate itself can sometimes satisfy the requirements for distribution preservation.

5. Key Results and Trade-offs

The paper validates its theoretical findings through Gaussian and Binary channel models:

  • Inner Bounds: Theorem 1 provides a mathematical characterization of the achievable region for communication rate, CR rate, and distortion.
  • Gaussian Trade-offs: Analysis of Gaussian channels reveals that higher correlation between the encoder's side information and the actual state reduces the need for Common Randomness while increasing the achievable communication rate.
  • Security Bounds: Theorem 3 establishes the first formal bounds for a "Secure CR-assisted distribution-preserving capacity-distortion region," quantifying the cost of privacy in ISAC systems.

6. Conclusion and Future Outlook

This research provides a rigorous foundation for secure and statistically accurate sensing in 6G. However, several avenues for future work remain. The current model assumes independent and identically distributed (i.i.d.) states, whereas real-world environments exhibit spatial and temporal correlations. Furthermore, while the paper establishes what is theoretically achievable (inner bounds), future research is needed to define the absolute theoretical limits (outer bounds) and develop low-latency, computationally efficient codes for real-time hardware implementation.

By treating randomness as a resource and prioritizing statistical fidelity, this framework ensures that ISAC can support the next generation of intelligent, environment-aware wireless services securely.

Summary updated

BibTeX
@inproceedings{chen2024distribution,
  author = {Yiqi Chen and Tobias J. Oechtering and Holger Boche and Mikael Skoglund and Yuan Luo},
  title = {{Distribution-Preserving Integrated Sensing and Communication with Secure Reconstruction}},
  booktitle = {ISIT},
  year = {2024},
  doi = {10.1109/isit57864.2024.10619214},
}
Figure