CAREER: Information theoretic methods for wireless networks

Carnegie Mellon University, Pittsburgh PA

Investigators

Abstract

Wireless networks have grown rapidly, resulting in increased productivity afforded by untethered mobile communications. The end-to-end communication problem for wireless networks includes physical layer (PHY), Medium Access Control (MAC), and routing issues. This effort aims to provide a joint solution to the end-to-end problem, encompassing all three areas, to maximize efficiency by attempting to unify the theories that support them, namely information theory, queuing and graph theory. Specifically, the MAC problem is being investigated for a variety of situations, incorporating diverse traffic characteristics, advanced physical layer technologies, and potential denial of service, and efficient protocols are being designed. The effect on routing of channel fading is being investigated. The novelty of the project is in the use of information theoretic techniques, such as entropy bounds and error exponents, to analyze MAC and routing, thus complementing queuing and graph theoretic analysis. This project aims to increase the efficiency of wireless networks, whose development is important for the continued growth of global productivity. The project will result in the understanding and design of efficient MAC and routing protocols, which explicitly incorporate PHY characteristics, and are capable of handling practical wireless channel impairments, arising in diverse wireless networks. The results will be of importance to the communications industry in designing high-end competitive products. Further, the project will develop graduate student expertise in information theoretical techniques for wireless networking, and will allow several more undergraduates and graduates to participate through instructional experiments.