Due to lack of fundamental understanding on how to share wireless media among spatially distributed users, the energy consumption and bandwidth efficiency of wireless networks remains defficient and highly suboptimal. Prior investigations of wireless channel access have followed two separate paths that reflect vastly different viewpoints; namely, the traditional information-theoretic approach assumes perfect user coordination and ignores the modularized network architecture, while the traditional network-theoretic approach largely focuses on access control protocols and ignores the impact of the physical layer. This project will bridge the gap of the classical theories by developing a theoretical foundation for channel access in distributed wireless systems. More specifically, it will extend classical information theory by developing a channel coding theory for physical layer distributed communication, where users do not jointly design channel codes. It will also extend classical network theory by developing a medium access control (MAC) framework for distributed networking, where physical layer properties, such as joint multiuser message decoding and flexible adaptation of communication parameters, are efficiently exploited at the link layer.
The project contains two parts which respectively address the physical and the data link layers of distributed wireless networks. In Part I of the project, the goal is to develop a rigorous coding theory to characterize the fundamental limitis of distributed communication systems. The coding theory will support extensive communication performance tradeoffs and structured coding schemes with low computational complexity. Part II of the project contains three steps. The objective of the first step is to characterize optimal link layer distributed channel sharing schemes, their fundamental properties and performance improvements over classical schemes. The objective of the second step is to develop a unified MAC framework to achieve asymptotic optimal channel sharing in distributed networks via the joint adaptation of communication parameters (e.g., rate, power, transmission probability). The objective of the third step is to develop MAC algorithms with fast convergence properties that ensure efficient network operation in transient environments.
By extending information theory to distributed communication models, the project will advance the integration of information and network theories, and significantly improve the energy and bandwidth efficiency of wireless systems. Unification of the two classical theories will also influence the way modern communication network subjects are taught in Higher Education and attract more talented students to this field of acute national importance.
