Multiple-Input Multiple-Output Gaussian Broadcast Channels with Common and Confidential Messages

Hung D. Ly, Tie Liu, Yingbin Liang

Abstract

This paper considers the problem of the multiple-input multiple-output (MIMO) Gaussian broadcast channel with two receivers (receivers 1 and 2) and two messages: a common message intended for both receivers and a confidential message intended only for receiver 1 but needing to be kept asymptotically perfectly secure from receiver 2. A matrix characterization of the secrecy capacity region is established via a channel enhancement argument. The enhanced channel is constructed by first splitting receiver 1 into two virtual receivers and then enhancing only the virtual receiver that decodes the confidential message. The secrecy capacity region of the enhanced channel is characterized using an extremal entropy inequality previously established for characterizing the capacity region of a degraded compound MIMO Gaussian broadcast channel.

Introduction

In modern wireless communication, security is traditionally managed at the upper layers of the protocol stack through cryptographic methods. However, Physical Layer Security (PLS) offers a complementary approach by leveraging the inherent physical characteristics of the wireless channel to ensure confidentiality. This paper, authored by Hung D. Ly, Tie Liu, and Yingbin Liang, provides a rigorous mathematical foundation for PLS in multi-antenna systems, specifically characterizing the secrecy capacity region for MIMO Gaussian broadcast channels.

System Model and Problem Formulation

The authors investigate a scenario where a transmitter equipped with multiple antennas communicates with two receivers. The transmission consists of two distinct data types:

  1. Common Message ($W_0$): Intended for simultaneous reception by both Receiver 1 and Receiver 2.
  2. Confidential Message ($W_1$): Intended exclusively for Receiver 1, with the requirement that it remains asymptotically perfectly secure from Receiver 2 (i.e., the mutual information between the message and the eavesdropper's signal vanishes as the block length $n$ goes to infinity).
System models for the general case (a) and the aligned case (b) of the MIMO Gaussian broadcast channel.

The challenge lies in determining the secrecy capacity region—the set of all achievable rate pairs $(R_0, R_1)$ for the common and confidential messages, respectively. While this region was previously understood for single-antenna (SISO) systems, the MIMO case introduces significant mathematical complexity due to the non-commutative nature of matrix interactions.

Methodology: Channel Enhancement

To overcome these complexities, the researchers employ a technique known as Channel Enhancement. This involves constructing an "enhanced" version of the original channel that is mathematically more tractable yet shares the same capacity region.

The proof strategy involves:

  • Receiver Splitting: Receiver 1 is conceptually split into two virtual receivers—one dedicated to decoding the common message and the other to the confidential message.
  • Targeted Enhancement: Only the virtual receiver intended for the confidential message is enhanced.
  • Extremal Entropy Inequalities: The authors utilize sophisticated inequalities (notably those developed by Weingarten et al.) to bound the entropy of the matrix-based signals, proving that the enhanced channel's capacity matches the original.

The transition from the "general case" to the "aligned case" (as shown in the figure above) is a critical step in this mathematical simplification, where channel matrices are effectively normalized into noise covariances.

Key Insights

  • Optimality of Gaussian Signaling: A primary finding is that Gaussian signaling remains optimal for achieving the secrecy capacity region in MIMO broadcast channels, paralleling results for standard broadcast channels without secrecy constraints.
  • The Cost of Secrecy: The results provide a quantitative measure of the "secrecy penalty." By comparing the secrecy region to the standard broadcast region, engineers can determine the exact trade-off between public broadcast rates and private message security.
  • Channel Alignment: The secrecy rate is highly sensitive to the alignment of the channel matrices. If the eavesdropper's channel (Receiver 2) is closely aligned with the legitimate user's channel (Receiver 1), the achievable secrecy rate decreases significantly.

Results

The core contribution of the paper is Theorem 1, which provides a precise matrix-based characterization of the secrecy capacity region under a matrix power constraint. This is further extended in Corollary 3 to a more practical average total power constraint, facilitating its application in wireless system design. Numerical simulations in the paper demonstrate that the secrecy capacity region is a convex shape, illustrating the inevitable drop in the confidential message rate as the common message rate increases.

Limitations and Future Work

While foundational, the study assumes Perfect Channel State Information (CSI) at the transmitter, a condition rarely met in real-world dynamic environments where CSI is often imperfect or delayed. Additionally, the model is limited to two receivers. Extending these results to K-user scenarios and accounting for fast-fading channels remain open challenges for future research in 5G and 6G network security.

Summary updated

Abstract

This paper considers the problem of the multiple-input multiple-output (MIMO) Gaussian broadcast channel with two receivers (receivers 1 and 2) and two messages: a common message intended for both receivers and a confidential message intended only for receiver 1 but needing to be kept asymptotically perfectly secure from receiver 2. A matrix characterization of the secrecy capacity region is established via a channel enhancement argument. The enhanced channel is constructed by first splitting receiver 1 into two virtual receivers and then enhancing only the virtual receiver that decodes the confidential message. The secrecy capacity region of the enhanced channel is characterized using an extremal entropy inequality previously established for characterizing the capacity region of a degraded compound MIMO Gaussian broadcast channel.
BibTeX
@article{ly2010multiple,
  author = {Hung D. Ly and Tie Liu and Yingbin Liang},
  title = {{Multiple-Input Multiple-Output Gaussian Broadcast Channels with Common and Confidential Messages}},
  journal = {IEEE Trans. Information Theory},
  year = {2010},
  url = {https://openreview.net/forum?id=P9CgdhS7IJ},
}
Figure