home · mobile · calendar · colloquia · 2008-2009 · 

Colloquium - Tandra

Spectrum Sharing via Cognitive Radios: Opportunities and Challenges
University of California, Berkeley

Under the current regulatory model of static frequency assignment, a great deal of spectrum remains underused. This seeming waste represents an opportunity for frequency-agile cognitive radios to improve performance. Understanding this opportunity forces us to take a closer look at the whole question of "regulatory overhead". Our technical results show that regulatory changes are required before research in this area can truly impact practice. Recent baby steps (whitespace ruling in TV bands on November 4, 2008) taken by the FCC demonstrate that the government is serious about change, but the problem is that we do not yet know what this change should entail or even how we would recognize the right answer if we saw it.

The full scope of the problem weaves together signal processing, wireless communications, information theory, policy, economics, and law in a nontrivial way. In this talk I focus on one fundamental aspect of this problem, namely sensing spectrum for empty bands that are not being used at their current time and location for their primary purpose. I show that real world uncertainties in the noise plus interference, and the fading process make spectrum sensing very challenging. In particular, I show that there exist fundamental limits called SNR walls, below which robust detection is impossible, regardless of how many samples we take. I consider a host of detection algorithms like non-coherent, coherent and cyclostationary feature detectors and show that they are all limited by these SNR walls, although some are more robust than others.

The results on limits of robust detection have significant policy implications. If we wish to reduce the regulatory overhead by allowing the flexibility of cognitive use in licensed bands, then we must pay an overhead in the form of additional infrastructure. Examples of such an infrastructure are: collaboration among radios, geo-location, explicit signaling from the primary system, etc. I conclude my talk by giving a unifying framework that enables a conceptual understanding of this "overhead" in the context of spectrum sensing.

Rahul Tandra is currently a PhD candidate in EECS at the University of California, Berkeley. He received the MS degree in EECS from the University of California, Berkeley in 2005, and the BTech degree in Electrical Engineering from the Indian Institute of Technology Bombay in 2003. He is a member of the Wireless Foundations Research Center and the Berkeley Wireless Research Center. In 2006, he worked as a summer intern in Qualcomm Corporate R&D, developing spectrum sensing algorithms for the IEEE 802.22 standard. His work on fundamental limits for signal detection got the best student paper award at the IEEE DySpAN conference in 2007. His research interests are wireless communications, signal processing and robust statistics. He is particularly interested in fundamental research questions in dynamic spectrum sharing.

Sponsored by the Interdisciplinary Telecommunications Program.

Department of Computer Science
University of Colorado Boulder
Boulder, CO 80309-0430 USA
May 5, 2012 (14:13)