CIF: Small: Optimizing Two-way Communications with Feedback

Purdue University, West Lafayette IN

Investigators

Abstract

To achieve high data rates over wireless communication channels, feedback has customarily been used either to learn the channel state information, or to request the re-transmission of a failed reception. Most studies of feedback have focused on a one-way perspective, meaning data travels in one direction, and feedback -- typically assumed to be perfect -- in the other. This project extends beyond perfect one-way feedback to study noisy feedback in two-way time-varying networks. In modern two-way networks, feedback travels over the same channel as the data in the opposite direction, resulting in a tradeoff between improving rates in one direction and feeding back information to improve rates in the other. This tradeoff is affected by the rate of channel time-variation: no variation enhances the value of the channel state information feedback, whereas rapid variation renders such feedback to the transmitter outdated. If re-transmissions are needed, the re-transmitted messages may be judiciously combined using methods similar to network coding, or they may be collaboratively re-transmitted by several users, leading to additional and interrelated throughput tradeoffs. This research will develop a foundation for the study of feedback in the context of two-way time-varying wireless Gaussian networks. The intellectual merit lies in proposing a unified framework that captures the key tradeoffs particular to two-way networks and the presence of different types of feedback, including quantized channel state information (Q-CSI), Automatic Repeat reQuest (ARQ), or extensions and combinations thereof. This research will rely on, and contribute to, communication and information theory. Given the ubiquity of two-way wireless communications, this research seeks to significantly increase the information rate of future wireless networks. The project comprises outreach components through mentoring students, integrating theory with practice, and extending the impact of the investigator's research and teaching to broader audiences.