Transcoding: A new strategy for relay channels
Chih-Chun Wang, David J. Love, Dennis O. Ogbe · 2017
The relay channel is a traditional information-theoretic problem which has important applications in the Internet of Things (IoT) and other future communication networks. In this work, we focus on the simplest possible relaying model: the so-called separated (or two-hop) relay channel where there is no direct link between the source and the destination. Previous work has shown that for this channel, the decode-and-forward (DF) relaying strategy is capacity-achieving under the assumption of asymptotic block lengths. In this paper, however, we are interested in the finite-delay regime where simpler sub-optimal techniques like amplify-and-forward (AF) can be used to avoid the need for buffering at the relay. We present a new strategy called transcoding which presents a tradeoff between the low-latency advantages of amplify-and-forward and the high-rate, high-latency decode-and-forward scheme. Our results indicate that our simple, intuitive transcoding schemes outperform traditional relaying schemes in the finite-delay regime.