Leveraging Channel Feedback for Secure Communication

49 concepts23 papers1 proposal

In the classic wiretap channel model, achieving information-theoretic security requires a fundamental physical advantage: the channel from the transmitter (Alice) to the legitimate receiver (Bob) must be cleaner or less degraded than the channel from Alice to the eavesdropper (Eve). If Eve enjoys a superior signal-to-noise ratio (SNR) or fewer packet drops, the traditional secrecy capacity drops to zero.

This paradigm shifts entirely when we introduce a feedback loop from Bob back to Alice. Channel feedback—whether it takes the form of continuous noisy signals, public acknowledgments, or delayed channel state information—acts as an equalizer. As surveyed by Bin Dai et al. in Enhancing physical layer security via channel feedback: a survey, integrating feedback allows Alice and Bob to establish positive secrecy rates even in "reversely degraded" scenarios where Eve's physical channel is strictly superior to Bob's.

Feedback introduces a dynamic source of shared uncertainty. Because Eve experiences independent noise or channel erasures, she cannot perfectly observe the interaction between Alice and Bob. This interaction allows the legitimate parties to distill common randomness or generate cryptographic keys on the fly, effectively blinding an eavesdropper regardless of her computational power.

Modulo-Additive Channels and Noisy Feedback

To understand the core mechanism, consider a full-duplex communication system operating over a modulo-additive channel. In The Wiretap Channel with Feedback: Encryption over the Channel, Lifeng Lai et al. demonstrate how Bob can actively generate a random sequence and transmit it back to Alice over a noisy feedback link.

Because Bob knows exactly what sequence he generated, he can perfectly subtract this "key" from the combined signal he receives from Alice in subsequent time slots. Eve, however, only observes a noisy mixture of Alice's message and Bob's feedback signal. In a follow-up work, Secrecy Capacity of the Wiretap Channel with Noisy Feedback, Lifeng Lai et al. show a counterintuitive result: a noisy feedback channel can sometimes be more effective for security than a perfectly noiseless one. The noise on the feedback path prevents Eve from perfectly capturing the key Bob is sending to Alice, ensuring that Bob maintains a private state that Eve cannot fully reconstruct. Through this mechanism, the system can achieve a secrecy capacity equal to the full standard capacity of the main channel, rendering the wiretapper completely irrelevant to throughput.

Building on this, Bin Dai et al. in An Improved Feedback Coding Scheme for the Wire-Tap Channel proposed an improved feedback coding scheme where the feedback signal serves a dual purpose: generating a secret key for encryption and producing "cooperative help information" to refine the encoding and decoding processes, expanding the achievable secrecy rate beyond previous degraded scenarios.

Exploiting Packet Erasures and ARQ Protocols

In packet-switched wireless networks, feedback often takes a discrete form: Automatic Repeat reQuest (ARQ) acknowledgments (ACK/NACK). Instead of treating packet drops as a pure reliability failure, feedback-enabled security frameworks repurpose them as a source of cryptographic entropy.

Consider a broadcast erasure channel where Alice transmits packets to Bob, and Eve listens in. Due to independent fading, some packets will be received by Bob but erased (dropped) for Eve. In Secret Communication over Broadcast Erasure Channels with State-feedback, László Czap et al. outline a two-phase scheme:

  1. Key Generation: Alice transmits random packets. Bob uses public feedback to acknowledge which packets he received successfully.
  2. Encrypted Transmission: Alice identifies the packets that Bob received but Eve missed. These exclusive packets are then used as a shared secret key (a one-time pad) to encrypt the actual private message.

Because the feedback is public, Eve knows which slots Bob received, but she lacks the content of those packets. This concept is implemented practically in wireless networks via protocols like ARQ-CROWN by Yara Abdallah et al. in Keys Through ARQ: Theory and Practice and evaluated in Wi-Fi and RFID settings by Mohamed Elsabagh et al. in ARQ security in Wi-Fi and RFID networks. By exploiting missed acknowledgments and the natural randomness of multipath fading, these protocols continuously update secret keys, blinding the eavesdropper to the evolving key state.

This approach scales efficiently to multi-hop networks. Athanasios Papadopoulos et al. in LP formulations for secrecy over erasure networks with feedback formulated linear programming (LP) frameworks to maximize secure communication rates over erasure networks with feedback. They demonstrate that public feedback allows nodes to dynamically identify and discard packets potentially intercepted by Eve, effectively starving the adversary of information. In Secret message capacity of a line network, Athanasios Papadopoulos et al. proved that this optimization can be achieved in polynomial time, even when intermediate nodes lack private randomness.

The role of feedback in time-varying channels is further explored by Bin Dai et al. in Finite State Markov Wiretap Channel with Delayed Feedback, where they demonstrate that delayed feedback of channel output and state information significantly boosts secrecy capacity by allowing legitimate parties to establish shared secrets hidden from an eavesdropper, even though feedback does not improve the capacity of standard non-secure channels. Similarly, Mohsen Bahrami et al. in Key agreement over a state-dependent 3-receiver broadcast channel show that a single round of limited, insecure public feedback expands the achievable private key capacity region by allowing distributed source coding techniques to reconcile key discrepancies caused by a lack of channel state information.

Topological Feedback and State Secrecy

The utility of feedback extends beyond point-to-point links into network topologies and cyber-physical systems.

In network graphs, the presence of cycles or undirected edges allows nodes to send information "backwards." Shaunak Mishra et al. in Using Feedback for Secrecy over Graphs demonstrated that these feedback loops allow nodes to establish common randomness without pre-shared keys. By routing random keys backward through network cycles, a node can provide a source with a secret pad that remains hidden from adversaries positioned exclusively on the forward paths, yielding higher secrecy rates than directed acyclic graphs.

In the context of control systems and remote state estimation, feedback can be used to protect the real-time state of a dynamical system. Anastasios Tsiamis et al. introduced "State-Secrecy Codes" (SSCs) for networked linear systems in State-Secrecy Codes for Networked Linear Systems. Here, a sensor transmits state estimates to a receiver. By utilizing an acknowledgment-based synchronization protocol detailed by the authors in An Information Matrix Approach for State Secrecy, the sensor subtracts a reference state known only to the authorized receiver based on its packet reception history. If an eavesdropper misses just a single packet, her lack of a synchronized reference point causes her estimation error to explode exponentially to the worst-case scenario, while the legitimate user maintains perfect tracking.

While direct feedback loops are powerful, security can also be enhanced via correlated side information from a helper node. As shown by Ravi Tandon et al. in Secure Source Coding With a Helper, when a sender has access to a helper's correlated side information, the security of the transmission significantly improves, allowing for greater uncertainty at the eavesdropper without sacrificing the data rate.

Theoretical Limits and Converses

While feedback significantly expands the secrecy region, it does not grant infinite capacity. The fundamental limits of these systems are bounded by the nature of the feedback channel itself.

For instance, if the feedback channel is entirely public, its utility is constrained because the eavesdropper monitors every return signal. Masahito Hayashi et al. established a strong converse for degraded wiretap channels with public feedback in Strong Converse for a Degraded Wiretap Channel via Active Hypothesis Testing using active hypothesis testing, proving that public feedback cannot arbitrarily increase capacity beyond strict information-theoretic thresholds.

However, when combined with delayed channel state information at the transmitter (CSIT) and output feedback in multi-antenna settings, distributed systems can achieve remarkable efficiency. Abdellatif Zaidi et al. showed in Secure Degrees of Freedom of MIMO X-Channels with Output Feedback and Delayed CSIT that distributed transmitters in a MIMO X-channel can match the secure transmission capacity of a centralized system, provided their feedback mechanisms are sufficiently robust.

Ultimately, feedback-based secret key generation and secure coding transform the physical layer from a passive conduit vulnerable to interception into an active, self-encrypting system that leverages its own imperfections for absolute confidentiality.

Key Foundational Work

Secrecy Capacity of the Wiretap Channel with Noisy Feedback (2007) — contributed 5 concept elements to this direction.
The Wiretap Channel With Feedback: Encryption Over the Channel (2007) — contributed 4 concept elements to this direction.
The Relay-Eavesdropper Channel: Cooperation for Secrecy (2008) — contributed 1 concept elements to this direction.
History & citationVersion 1 · 2026-10-04

Current · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_feedback_based_secret_key_generation_b85a1df5,
  title = {feedback-based secret key generation},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/feedback-based-secret-key-generation/rdr_b85a1df5b7b04ae0b20d8629ad71f78d},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-20},
}

This exact version (version 1, frozen text):

@misc{lacuna_feedback_based_secret_key_generation_b85a1df5_v1,
  title = {feedback-based secret key generation},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/feedback-based-secret-key-generation/rdr_b85a1df5b7b04ae0b20d8629ad71f78d/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-20},
}
About

These directions overlap with this one through shared papers.

See something wrong? Report a mistake