MIMO Gaussian Broadcast Channels with Confidential and Common Messages

This paper considers the problem of secret communication over a two-receiver multiple-input multiple-output (MIMO) Gaussian broadcast channel. The transmitter has two independent, confidential messages and a common message. Each of the confidential messages is intended for one of the receivers but needs to be kept perfectly secret from the other, and the common message is intended for both receivers. It is shown that a natural scheme that combines secret dirty-paper coding with Gaussian superposition coding achieves the secrecy capacity region. To prove this result, a channel-enhancement approach and an extremal entropy inequality of Weingarten et al. are used.

💡 Research Summary

The paper studies a two‑receiver multiple‑input multiple‑output (MIMO) Gaussian broadcast channel in which the transmitter must deliver three independent streams: a common message intended for both receivers and two confidential messages, each intended for only one receiver while being kept perfectly secret from the other. The model assumes fixed channel matrices (\mathbf{H}_1) and (\mathbf{H}_2), additive white Gaussian noise at each receiver, and a total transmit‑power constraint (P). Secrecy is defined in the strong information‑theoretic sense: for each confidential message (W_i) ((i=1,2)), the mutual information with the unintended receiver’s observation must vanish as blocklength grows.

The main contribution is a complete characterization of the secrecy capacity region (\mathcal{C}) for this setting. The authors propose a natural coding scheme that combines secret dirty‑paper coding (S‑DPC) for the confidential streams with Gaussian superposition coding for the common stream. Concretely, the transmitted signal is decomposed as (\mathbf{X} = \mathbf{X}_0 + \mathbf{X}_1 + \mathbf{X}_2), where (\mathbf{X}_0) carries the common message, and (\mathbf{X}_i) ((i=1,2)) are generated by S‑DPC that pre‑cancels the interference caused by the other confidential stream while embedding randomization to guarantee secrecy. Power is allocated through positive‑semidefinite covariance matrices (\mathbf{K}_0,\mathbf{K}_1,\mathbf{K}_2) satisfying (\mathbf{K}_0+\mathbf{K}_1+\mathbf{K}_2 \preceq P\mathbf{I}). Each receiver first decodes the common layer, subtracts it, and then decodes its own confidential layer using the DPC structure; the unintended receiver sees only a statistically independent Gaussian interference, ensuring the secrecy constraint.

To prove optimality, the authors employ a channel‑enhancement technique. They construct an auxiliary “enhanced” broadcast channel with reduced noise covariances, which dominates the original channel in the sense that any achievable rate tuple for the original channel is also achievable for the enhanced one. For the enhanced channel, they invoke the extremal entropy inequality of Weingarten, Steinberg, and Shamai (2006), which states that under a covariance constraint the Gaussian distribution maximizes differential entropy. This inequality yields tight upper bounds on the achievable rates for each layer. The resulting bounds are expressed in closed‑form as logarithmic determinants involving (\mathbf{H}_1,\mathbf{H}_2) and the covariance matrices (\mathbf{K}_i). By carefully choosing (\mathbf{K}_i) the proposed S‑DPC + superposition scheme meets these bounds with equality, establishing that the scheme attains the secrecy capacity region.

The paper also discusses several special cases. When the common message is omitted, the region reduces to the known secrecy capacity region for the MIMO broadcast channel with two confidential messages. When each node has a single antenna (SISO), the results collapse to the classic scalar Gaussian broadcast secrecy region. These consistency checks validate the general formulas.

From a practical viewpoint, the authors note that while exact S‑DPC is computationally intensive, linear precoding approximations and lattice‑based implementations can approach the theoretical performance. The superposition of the common layer is compatible with existing multi‑user OFDM or CDMA frameworks, making the scheme attractive for secure wireless standards.

In summary, the paper delivers a rigorous, complete description of the secrecy capacity region for MIMO Gaussian broadcast channels with both confidential and common messages, and demonstrates that a combination of secret dirty‑paper coding and Gaussian superposition coding is both sufficient and necessary to achieve this region. The methodological tools—channel enhancement and the extremal entropy inequality—provide a powerful template for future secrecy analyses in more complex multi‑user MIMO networks.