| Andrea Goldsmith | Ramesh Johari | |
|---|---|---|
| Professor, Electrical Engineering | Assistant Professor, Management Science and Engineering | |
| E-mail: andrea@ee.stanford.edu | and by courtesy, Electrical Engineering and Computer Science | |
| E-mail: ramesh.johari@stanford.edu |
The FCC today maintains relatively tight control of spectrum access, through a variety of regulations and licensing programs. However, this is an artifact of the past more than a harbinger of the future. We propose a ``clean slate'' design of wireless spectrum allocation, to deal with a future where multiple devices will share resources across broad ranges of space, time, and frequency. Our project aims for both design recommendations for wireless devices operating in a competitive environment, as well as protocol suggestions to encourage cooperation among opportunistic wireless devices. Our main methodological tools are derived from game theory, distributed control, and wireless system design.
There has been tremendous growth over the last decade in devices that
access the Internet via wireless technology. However, wireless spectrum
regulations have not kept pace with the growth in wireless end hosts;
thus it appears as though spectrum in scarce, when in reality this
scarcity arises due to access restrictions
imposed by the FCC. As suggested by an FCC task force in 2002, the
future Internet needs to have a more flexible wireless resource
allocation scheme. Our research will enable devices to find and utilize
spectral holes across a wide range of space, time, and frequency; such
research is ultimately vital to ensuring that spectrum availability for
the future Internet is commensurate with the growth in demand for
wireless services.
Most work in the area of dynamic spectrum allocation is based on idealized assumptions and/or ad-hoc techniques. The proposed research aims to develop a concrete formulation of the optimization problems associated with distributed dynamic spectrum allocation and also obtained some preliminary results indicating (a) the potential spectral efficiency gains of dynamic spectrum allocation; (b) the most promising techniques to use; and (c) the most significant technical challenges involved in applying these techniques in practice (e.g. computational complexity, overhead, measuring interference, radio hardware requirements).