Listen-and-Talk: Protocol Design and Analysis for Full-duplex Cognitive Radio Networks

Yun Liao, Tianyu Wang, Lingyang Song, Zhu Han

Introduction

In traditional cognitive radio networks, secondary users (SUs) typically operate in a half-duplex mode using a "Listen-Before-Talk" (LBT) protocol. While functional, LBT suffers from two major inefficiencies: spectrum waste during dedicated sensing periods and a high risk of collisions if a primary user (PU) returns while the SU is transmitting. This paper proposes the "Listen-and-Talk" (LAT) protocol, which leverages full-duplex (FD) hardware to allow SUs to sense and transmit simultaneously.

The Listen-and-Talk (LAT) Protocol

The LAT protocol utilizes a dual-antenna configuration at the SU: one antenna is dedicated to continuous spectrum sensing, while the other handles data transmission. This architecture enables the SU to maintain constant awareness of the channel state.

System model and timing diagram of the LAT protocol, illustrating the interaction between PU traffic and SU sensing/transmission.

As shown in the timing diagram, the SU can detect the PU's arrival or departure in real-time. However, this simultaneous operation introduces Residual Self-Interference (RSI)—noise from the SU's own transmission that leaks into the sensing antenna. To mitigate this, the authors employ adaptive energy detection with two distinct thresholds: $\epsilon_0$ for when the SU is silent and $\epsilon_1$ for when it is active.

Mathematical Modeling and Analysis

To evaluate the protocol's performance, the researchers modeled the system as a Discrete-Time Markov Chain (DTMC). This model tracks transitions between four fundamental states: PU only, SU only, collision, and idle (spectrum waste).

State transition diagram for the joint PU and SU activity, used to derive closed-form expressions for throughput and collision ratios.

By analyzing these transitions, the authors derived analytical expressions for the collision ratio and spectrum waste ratio, enabling the identification of an optimal transmit power that maximizes throughput while maintaining required sensing accuracy.

Key Insights and Results

The study reveals a critical Power-Throughput Tradeoff. Unlike traditional systems where higher power generally increases throughput, in LAT, excessive transmit power elevates RSI. This "blinds" the SU's sensing mechanism, leading to a higher false alarm rate and a subsequent drop in throughput. The authors identified a "sweet spot" for transmit power that balances signal strength with sensing reliability.

Key performance improvements include:

  • Throughput Gains: LAT significantly outperforms LBT by reclaiming the time previously lost to "quiet" sensing periods.
  • Reduced Collision Duration: Because sensing is continuous, the SU can react to a PU's return almost immediately, reducing average collision lengths to approximately half a time slot ($T/2$), compared to the full sub-slots typical of LBT.
  • Efficiency Limits: The analysis shows that under severe RSI, the system's efficiency tends toward 0.5 as the SU's sensing becomes unreliable, causing it to toggle on and off every other slot.

Conclusion and Future Work

The LAT protocol represents a significant step forward for cognitive radio efficiency. By integrating full-duplex sensing, it minimizes the "blind spots" inherent in LBT. While the theoretical results are promising, practical implementation faces challenges such as the hardware complexity of high-performance RSI cancellation in low-power devices. Future research will focus on extending this model to multi-user and multi-channel environments to address inter-SU interference in complex network topologies.

Summary updated

BibTeX
@article{liao2017listen,
  author = {Yun Liao and Tianyu Wang and Lingyang Song and Zhu Han},
  title = {{Listen-and-Talk: Protocol Design and Analysis for Full-duplex Cognitive Radio Networks}},
  journal = {IEEE Trans. Vehicular Technology},
  year = {2017},
  doi = {10.1109/tvt.2016.2535483},
}
Figure