Abstract
We consider a three-terminal state-dependent relay channel with the channel state noncausally available at only the relay. Such a model may be useful for designing cooperative wireless networks with some terminals equipped with cognition capabilities, i.e., the relay in our setup. In the discrete memoryless (DM) case, we establish lower and upper bounds on channel capacity. The lower bound is obtained by a coding scheme at the relay that uses a combination of codeword splitting, Gel'fand-Pinsker binning, and decode-and-forward relaying. The upper bound improves upon that obtained by assuming that the channel state is available at the source, the relay, and the destination. For the Gaussian case, we also derive lower and upper bounds on the capacity. The lower bound is obtained by a coding scheme at the relay that uses a combination of codeword splitting, generalized dirty paper coding, and decode-and-forward relaying; the upper bound is also better than that obtained by assuming that the channel state is available at the source, the relay, and the destination. In the case of degraded Gaussian channels, the lower bound meets with the upper bound for some special cases, and, so, the capacity is obtained for these cases. Furthermore, in the Gaussian case, we also extend the results to the case in which the relay operates in a half-duplex mode.
Introduction: The Challenge of Information Asymmetry
In modern wireless communications, "channel state" typically refers to interference or environmental factors that degrade signal quality. Traditional information theory often assumes either universal knowledge of this state or knowledge restricted to the transmitter. However, emerging "cognitive" networks introduce a more complex scenario: information asymmetry.
Consider a system where a sophisticated relay node can sense or predict interference (the state) non-causally, while the original source and the final destination remain "blind" to it. The fundamental question is how the relay can leverage its private knowledge of the interference to assist the source, especially since the source cannot adapt its signal to the state.
System Model and Methods
The paper investigates a three-terminal relay channel where the state $S^n$ is known non-causally only at the relay. This setup is depicted in the system model below.
To address the tension between cooperation and state cancellation, the authors introduce several key innovations:
- Codeword Splitting: This is the core technical contribution. The relay's signal is split into two parts: one dedicated to cooperating with the source to transmit the message, and another dedicated to neutralizing the channel state.
- Advanced Binning and Coding: For discrete memoryless channels, the relay employs Gel’fand-Pinsker binning to "pre-cancel" interference. In Gaussian environments, this evolves into a generalized version of Costa’s "Dirty Paper Coding" (DPC).
- Relaying Strategies: The authors primarily use Decode-and-Forward (DF), where the relay decodes the source's message before re-encoding. They also explore Partial DF (rate-splitting) for broader applicability.
- Tighter Upper Bounds: Moving beyond the standard cut-set bound, the authors derive a non-trivial upper bound that accounts for the rate penalty incurred by the source's ignorance of the channel state.
Key Insights and Results
The research yields several significant findings regarding the capacity of these asymmetric channels:
- Resolving Cooperation Tension: Codeword splitting effectively allows the relay to maximize cooperative gain while simultaneously mitigating interference, solving a major design trade-off.
- Quantifying the "Cost of Ignorance": The paper provides a mathematical framework to quantify the rate loss caused by the source not knowing the state, leading to more realistic performance bounds.
- Capacity for Degraded Channels: For specific "degraded" Gaussian channels—where one link is a noisier version of another—the proposed lower and upper bounds coincide, establishing the exact capacity.
- Multi-hop Optimality: In scenarios with extremely strong interference, a two-hop strategy (canceling interference during reception and using DPC during transmission) can achieve capacity.
- Practical Extensions: Numerical simulations show that the gap between the achievable rate and the upper bound is small for general Gaussian channels. Furthermore, the results hold for half-duplex relays, which are more common in practical hardware.
Limitations and Future Work
While the paper makes significant strides, certain limitations remain. The reliance on Decode-and-Forward (DF) means the results are less applicable when the source-to-relay link is weak; in such cases, "Compress-and-Forward" might be superior. Additionally, a gap between bounds persists at low Signal-to-Noise Ratios (SNR), particularly on the "broadcast" side of the relay. Finally, while the proposed coding schemes are theoretically optimal, their real-time implementation complexity remains a hurdle for future practical development.
Summary updated