Fundamentals of cooperation in modern communication networks

点击翻译按钮获取中文摘要

User cooperation is essential for the proper function of any communications network. For example, future networks such as the Internet of Things will be decentralized and ad-hoc, allowing a multitude of mobile terminals to continuously join and leave. Many of these terminals will be small-scale devices whose resources (battery, antennas, etc.) are too limited to maintain communication on their own. Such devices depend on other devices in their proximity to help carry their signal to its destination. A cooperative network thus provides the necessary infrastructure and intelligence for interactions to materialize.This project aims to contribute to three major areas of cooperative communications: (1) network information theory; (2) secure cooperation via physical layer security (PLS); and (3) codes for cooperation and security. Our work on information theory will focus on fundamental understanding, including compression and binning strategies, interactive communication, feedback, and multiuser duality. Our work on security is motivated by the observation that cooperation inherently conflicts with privacy and security due to the requirement for data sharing. We combine cooperation and security via physical layer security (PLS) where the key idea is to exploitthe noise and variations of the communications channel as sources of randomness. Information theory usually uses random coding arguments to gain insight. We are also interested in designing practical codes for cooperation and security. We will study network coding for reliability via vector-linear and quasi-linear codes, and for distributed storage systems where network nodes cooperate to recover from disk failures. Here the concepts of locality, availability, and private information retrieval play central roles. Finally, wewill study randomness to enable secrecy and stealth. Stealth coding is based on resolvability and distribution matching that has applications to shaping for coded modulation, steganography, and PLS. We will develop the theory of distribution matching, and design good codes, encoders, and decoders for short and moderate block lengths.