[Paper Review] On the Fading Paper Achievable Region of the Fading MIMO Broadcast Channel

Amir Bennatan, David Burshtein|ArXiv.org|Oct 5, 2006

Cooperative Communication and Network Coding20 references4 citations

TL;DR

This paper introduces a 'fading paper' transmission scheme for the ergodic fading MIMO broadcast channel with partial channel state information at the transmitter. Using convex optimization, it proves that Gaussian signaling achieves the linear-assignment capacity and demonstrates significant rate gains over non-adaptive dirty paper coding by exploiting channel state information, with numerical results showing close approach to the Sato upper bound.

ABSTRACT

We consider transmission over the ergodic fading multi-antenna broadcast (MIMO-BC) channel with partial channel state information at the transmitter and full information at the receiver. Over the equivalent {\it non}-fading channel, capacity has recently been shown to be achievable using transmission schemes that were designed for the ``dirty paper'' channel. We focus on a similar ``fading paper'' model. The evaluation of the fading paper capacity is difficult to obtain. We confine ourselves to the {\it linear-assignment} capacity, which we define, and use convex analysis methods to prove that its maximizing distribution is Gaussian. We compare our fading-paper transmission to an application of dirty paper coding that ignores the partial state information and assumes the channel is fixed at the average fade. We show that a gain is easily achieved by appropriately exploiting the information. We also consider a cooperative upper bound on the sum-rate capacity as suggested by Sato. We present a numeric example that indicates that our scheme is capable of realizing much of this upper bound.

Motivation & Objective

• To develop an achievable rate region for the fading MIMO broadcast channel under partial channel state information at the transmitter.
• To define and analyze the linear-assignment capacity as a tractable proxy for the full fading paper capacity.
• To prove that Gaussian signaling maximizes the linear-assignment capacity using convex analysis methods.
• To compare the proposed scheme with a non-adaptive dirty paper coding approach that assumes average channel state, quantifying the gain from exploiting partial CSI.
• To evaluate the performance of the scheme relative to a cooperative upper bound (Sato bound) and assess its tightness.

Proposed method

• Introduces a 'fading paper' model as a generalization of Costa's dirty paper coding to time-varying MIMO broadcast channels.
• Defines the linear-assignment capacity as the maximum mutual information under a linear power allocation constraint across channel realizations.
• Uses convex optimization techniques to prove that the optimal input covariance matrix for the linear-assignment capacity is Gaussian.
• Models the channel with four discrete fading states, each with a probability distribution constrained to preserve marginal user statistics.
• Introduces correlated virtual noise and channel matrices across states to model dependencies while preserving individual user statistics.
• Numerically evaluates the achievable rate region using semidefinite programming and computes a cooperative upper bound via Sato's method by optimizing over correlation parameters.

Experimental results

Research questions

• RQ1 Can a fading paper approach achieve a significant rate gain over non-adaptive dirty paper coding in fading MIMO broadcast channels with partial CSI?
• RQ2 Is the linear-assignment capacity, a tractable proxy for the full fading paper capacity, maximized by Gaussian signaling?
• RQ3 How close can the proposed fading paper scheme come to the Sato cooperative upper bound on the sum-rate capacity?
• RQ4 What is the impact of correlation between channel states and noise on the achievable rate region?
• RQ5 Does exploiting partial channel state information at the transmitter lead to a non-trivial performance gain over schemes assuming fixed average channel conditions?

Key findings

• The linear-assignment capacity is maximized by a Gaussian input distribution, proven via convex analysis and duality methods.
• A substantial rate gain is achieved by the proposed fading paper scheme over a non-adaptive dirty paper coding approach that ignores channel state variations and assumes average fading.
• Numerical results show that the proposed scheme achieves a sum-rate very close to the Sato cooperative upper bound, particularly when correlation parameters are optimized.
• The tightest Sato bound in the numerical evaluation was obtained with α=0 and ρ₁=ρ₂=ρ₃=ρ₄=0.3, indicating limited need for complex correlation structures.
• The performance gap between the proposed scheme and the Sato bound is small, suggesting the scheme is near-optimal in the studied setup.
• The analysis confirms that even partial channel state information at the transmitter enables significant spectral efficiency gains in fading MIMO broadcast channels.