Many electronic communication systems are bidirectional: they can transmit information in both directions. At any given time, however, the system may only be actively transmitting information in one direction. In this case, the reverse link, which would otherwise be idle, can be used to facilitate transmission in the forward direction. While this fact has long been recognized, it is not well understood how to best employ feedback to maximize the rate and minimize complexity for the forward link. This project will develop practical schemes that improve communication system performance by using feedback links and mathematically characterizing optimal methods for using feedback links. The improved understanding of how to use feedback should improve wireless communication technologies for cellular systems and the internet of things. The project will broaden participation in computing by offering research opportunities to women undergraduates in electrical engineering, with the goal of increasing the number of women who attend graduate programs in the field. It will also develop a publicly-accessible service that provides channels "in the cloud" for digital communication courses and research projects.
Historically, research in information and coding theory has focused on idealized communication scenarios with either perfect feedback or none at all. Many practical systems, however, have feedback links that are available but imperfect. This project will examine how to best use communication channels with a limited amount of feedback to achieve fast and reliable transmission of information. The project will provide clear theoretical guidance about how much feedback information is useful and how the transmitter should use feedback most effectively to adapt its encoding strategy. The project will provide bounds and approximations on the communication rates and delays achievable with a limited number of feedback transmissions. Leveraging a novel view of feedback coding as a form of joint source-channel coding, the project will address open questions concerning the second-order coding rate and "moderate deviations" performance with feedback. Finally, the project will explore how to improve delay-reliability tradeoffs achievable with feedback in the scenario where an information source, such as a sensor, produces information continuously, by combining the new information with the old information to encode immediately instead of waiting to buffer long blocks of data to encode with traditional block codes.
This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
