CAREER: Opportunistic Communication: A Design Paradigm for Wireless Systems

University Of Illinois At Urbana-Champaign, Urbana IL

Investigators

Abstract

Demand for wireless communication services has been growing much faster than additional resources (such as bandwidth) are allocated. Besides scarce available bandwidth, wireless system designers have to deal with several unique challenges, not commonly found in wireline networks, such as interference caused by transmissions and user specific time and frequency varying channels. Managing the limited resources through a harmonious design of the physical and networking layers is a central problem in the design of wireless systems and a very relevant one. The intense research activity over the past decade addressing this issue can be roughly dichotomized into two parts: the physical layer work focusing mostly on point to point communication and the networking layer work of resource allocation using an abstraction of the physical layer. Further, engineering design of wireless systems, for the most part, has centered around making individual point to point links reliable and developing a networking layer using an appropriate model of the physical layer. This project takes a fundamental physical layer view of the wireless network communication problems and sheds insight into novel design techniques at both the physical and the networking layer, yielding a significant boost to the overall system capacity. One novel viewpoint is multiuser diversity, a form of diversity inherent in a wireless network. In a system with many users whose channels vary independently, there is likely to be a user whose channel is near its peak at any one time. By scheduling transmissions only to users with the best channel, multiuser diversity gain is harnessed. This form of "opportunistic communication" provides a sharp boost to the system throughput. Through the lens of multiuser diversity, channel fluctuations are preferable, in sharp contrast to the point to point communication view. This project conducts a systematic study of the role of multiuser diversity: schemes to artificially induce and harness it. Other specific problems studied include a fundamental view of the multiple antenna downlink (broadcast) channel and an ad hoc networking model. Teaching is an integral part of this project and is partly carried out through the development of four courses.