SURA: Secure Unsourced Random Access

Mohammad Javad Ahmadi, Rafael F. Schaefer, H. Vincent PoorarXiv 2025 · DOI · arXiv

Abstract

Massive machine-type communications (mMTC) in 6G require efficient and secure access protocols. While Unsourced Random Access (URA) provides high scalability for grant-free communication, it lacks inherent security against eavesdropping. This paper proposes SURA (Secure Unsourced Random Access), a framework that integrates Physical Layer Security (PLS) into URA. By utilizing base station feedback for key generation and applying artificial noise to mask key information, SURA achieves data confidentiality with zero additional signaling overhead. Simulation results demonstrate a controllable trade-off between secrecy and reliability, proving SURA's viability for secure, large-scale IoT deployments.

Introduction

As we transition toward 6G, millions of low-power IoT devices are expected to connect to the network simultaneously. To handle this density, researchers developed Unsourced Random Access (URA), where devices transmit data in a shared resource pool without prior identification or grants. While URA is highly efficient for mMTC, its open nature makes it inherently vulnerable to eavesdropping. Traditional encryption methods like AES are often too computationally expensive for battery-powered sensors, and managing secret keys for millions of anonymous devices is a significant logistical challenge. SURA addresses these issues by providing lightweight security that utilizes the physics of the wireless channel itself.

The SURA Framework

The SURA scheme employs Physical Layer Security (PLS) to ensure privacy through three innovative steps:

  1. Feedback-Based Key Generation: Users listen to existing feedback signals broadcast by the Base Station (BS) for sensing or synchronization. Since every user perceives this signal through a unique, location-dependent wireless channel, it acts as a private observation that an eavesdropper (Eve) cannot perfectly replicate.
  2. Three-Segment Signal Structure: Each transmission is divided into a Pilot Segment (for channel estimation), a Polar Segment (containing encrypted data), and a Key Segment. The Key Segment contains parity bits of the secret key masked by Artificial Noise (AN) derived from the user's private channel.
  3. Syndrome-Based Reconciliation: To conserve bandwidth, users transmit only the parity bits (syndromes) of the secret key. The BS, leveraging its own estimate of the user's channel, uses these bits to reconstruct the full key via Slepian-Wolf decoding.
  4. Iterative Receiver Design: The BS employs an iterative algorithm involving Orthogonal Matching Pursuit (OMP) and Successive Interference Cancellation (SIC) to untangle signals from multiple simultaneous users.

Performance and Trade-offs

The performance of SURA was evaluated using Monte Carlo simulations under standard Rayleigh fading conditions, testing up to 100 active users ($K_a$) sharing the same time-frequency slots. A central finding of the study is the direct trade-off between security and reliability.

Trade-off between Per-User Probability of Error (PUPE) and Equivocation Rate for varying numbers of active users ($K_a$).

As illustrated in the figure above, increasing the ratio of Artificial Noise ($P_a/P_k$) enhances security—measured by the normalized equivocation rate, or Eve's uncertainty—but simultaneously leads to a higher Per-User Probability of Error (PUPE) at the BS. This "secrecy gap" remains robust even as the number of active users increases, demonstrating the system's scalability. The analysis assumes a worst-case scenario where the eavesdropper has perfectly cancelled out interference from other users, yet the system maintains high levels of secrecy.

Conclusion and Future Work

SURA demonstrates that meaningful secrecy can be achieved in massive IoT networks by "recycling" existing system features like feedback signals. This approach adds zero signaling overhead while maintaining the low-latency benefits of URA. Future work will address active attackers (jamming or spoofing), the impact of physical proximity between users and eavesdroppers, and the performance of these techniques in multi-antenna (MIMO) configurations where perfect channel estimation may not be possible.

Summary updated

BibTeX
@misc{ahmadi2025sura,
  author = {Mohammad Javad Ahmadi and Rafael F. Schaefer and H. Vincent Poor},
  title = {{SURA: Secure Unsourced Random Access}},
  year = {2025},
  eprint = {2512.09104},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2512.09104},
}
Figure