On the Relationship Between the Multi-antenna Secrecy Communications and Cognitive Radio Communications

Lan Zhang, Rui Zhang, Ying-Chang Liang, Yan Xin, Shuguang Cui

Abstract

This paper establishes a fundamental mathematical bridge between two traditionally distinct fields of wireless communications: Physical Layer Secrecy and Cognitive Radio (CR). By demonstrating that the optimal transmit strategy for a multi-antenna secrecy channel is equivalent to finding the capacity of a cognitive radio channel under specific interference constraints, the authors provide a framework to solve previously intractable non-convex optimization problems. This duality enables the calculation of exact secrecy capacities for single-antenna eavesdroppers and provides tight performance bounds for multi-antenna scenarios.

Introduction

In modern wireless communications, Multiple-Input Multiple-Output (MIMO) systems are essential for achieving high data rates. However, the broadcast nature of wireless signals makes them inherently vulnerable to interception. Secrecy Capacity—defined as the maximum rate at which information can be transmitted to a legitimate receiver while ensuring an eavesdropper gains zero information—is the primary metric for physical layer security.

Mathematically, determining secrecy capacity involves a complex "max-min" optimization: maximizing the legitimate user's information rate while minimizing the leakage to eavesdroppers. In environments with multiple eavesdroppers, this problem becomes non-convex, characterized by multiple local optima that frustrate standard optimization techniques. Historically, finding the global optimum for the general multi-eavesdropper case was computationally prohibitive.

The Secrecy-Cognitive Radio Duality

The core contribution of this research is the discovery of a "dual" relationship between secrecy systems and Cognitive Radio. In a CR network, a Secondary User (SU) is permitted to transmit only if its signal does not exceed a specific "Interference Temperature" (IT) at the Primary User (PU).

The authors observed that an eavesdropper in a secrecy system behaves mathematically like a PU in a CR system. In both instances, the transmitter must constrain the signal power reaching a specific node—either to prevent interference (CR) or to prevent interception (Secrecy). The paper proves that for any MIMO secrecy problem, there exists a corresponding set of IT constraints in an auxiliary CR model that yields the identical optimal transmit covariance matrix. This mapping allows researchers to apply well-understood convex optimization tools from CR research to the field of information-theoretic security.

Algorithmic Framework

By identifying that the secrecy objective function is quasi-concave with respect to interference constraints, the authors developed efficient algorithms to find global solutions:

  1. General Case Bisection: The authors proposed a bisection search method that decomposes the non-convex secrecy problem into a sequence of convex feasibility problems. These can be solved efficiently using interior-point or ellipsoid methods.
  2. Simplified Single-Eavesdropper Case: A streamlined version of the algorithm was developed for scenarios with only one eavesdropper, significantly reducing computational overhead.
  3. MISO Optimization: For Multiple-Input Single-Output (MISO) systems (where the receiver has a single antenna), the authors demonstrated that the problem simplifies into a Semi-Definite Programming (SDP) task, which is computationally very fast and always results in a simple beamforming strategy (a rank-one matrix).

Key Results and Performance

The proposed framework was validated through numerical simulations using independent Circularly Symmetric Complex Gaussian (CSCG) channel models. Key findings include:

  • Superiority over P-SVD: The algorithms consistently outperformed the Projected-Channel Singular Value Decomposition (P-SVD) baseline. While P-SVD attempts to completely null the signal in the direction of eavesdroppers, the authors' optimal strategy allows for minimal leakage if it disproportionately benefits the legitimate receiver's rate.
  • Exact Capacity for Single-Antenna Eavesdroppers: The paper provides the first efficient method to calculate the exact secrecy capacity for any number of single-antenna eavesdroppers.
  • Tight Bounds for MIMO Eavesdroppers: In the more challenging case where eavesdroppers possess multiple antennas and can perform joint processing, the authors established upper and lower bounds that are significantly tighter than previous estimates.

Limitations and Future Directions

While the paper represents a significant theoretical leap, several practical challenges remain. The model assumes perfect Channel State Information (CSI) for all eavesdroppers. In practice, eavesdroppers are often passive, making their channel gains difficult to estimate. Furthermore, while the bisection method is "efficient" in a complexity class sense, the iterative nature of solving multiple convex problems may still pose latency challenges for real-time, highly dynamic wireless links. Future research continues to explore robust beamforming techniques that account for CSI uncertainty and further close the gap in multi-antenna eavesdropper scenarios.

Summary updated

BibTeX
@inproceedings{zhang2009relationship,
  author = {Lan Zhang and Rui Zhang and Ying-Chang Liang and Yan Xin and Shuguang Cui},
  title = {{On the Relationship Between the Multi-antenna Secrecy Communications and Cognitive Radio Communications}},
  booktitle = {Allerton},
  year = {2009},
  doi = {10.1109/allerton.2009.5394527},
}
Figure