Skip to main content - access key m.
Skip to main navigation - access key n.

Wed. Dec. 1, 2010, 1 p.m.
TI Auditorium
(ECSS Room 2.102)

 

 

 

 

 

 

 

 

 

 

 EE seminar series

“Interference in General Random Wireless Networks”
Dr. Radha Ganti, UT Austin

Sponsored by the Dallas Chapter of the IEEE Signal Processing Society

Abstract
The spatial structure of node locations is critical in determining the mutual interference and hence the performance of a wireless network. A common assumption is the independence of transmitter locations, and spatial Poisson point process (PPP) is ubiquitously used to model the node locations. While in many cases a PPP model leads to analytical tractability, it precludes intelligent scheduling of transmitters which introduce correlations. To optimize the spatial packing of transmitters, we require new mathematical tools to analyze and understand the performance of general spatial networks, and we cannot always resort to the comfortable assumption of a Poisson distribution of nodes. This talk will introduce new tools and techniques from stochastic geometry and point process theory to analyze functions of interference in non-Poisson wireless networks. The first part will emphasize two important metrics: outage probability and transmission capacity (TC). A new technique to obtain outage probability of any wireless network with any arbitrary MAC protocol for a low density of transmitters will be introduced. It can be shown that the outage and the TC of any network can be completely characterized by two parameters: the spatial contention parameter and the interference scaling exponent. These parameters can easily be computed, as shall be illustrated using examples of cognitive, femtocell and CSMA networks. The second part of the talk will emphasize more recent work that generalizes the low-density outage results to any density and any function of interference. A new technique of factorial moment expansion that provides a Taylor-series of any arbitrary function of interference will be presented.

Bio
Dr. Radha Krishna Ganti is a postdoctoral researcher in the Wireless Networking and Communications Group at UT Austin. He received his PhD in EE from the University of Notre Dame in 2009. His doctoral work focused on the spatial analysis of interference networks using tools from stochastic geometry. He is the co-author of the monograph “Interference in Large Wireless Networks” (NOW Publishers, 2008).