Mini-tutorial #2: Foundations of Sensor Nets
Greg Pottie – UCLA

Title: Multi-terminal Information Theory Problems in Sensor Networks

Abstract:

Sensor networks are fundamentally concerned with the extraction of information about the physical world, and its conveyance to end-users. The information must be gathered according to some fidelity criterion, while meeting resource contraints such as energy or bandwidth. The sensing nodes may collaborate in communicating and performing the necessary signal processing. We show that such networks are scalable, that is, performance improves as the density increases, in contrast to ad hoc networks in which the transport capacity per node goes to zero. We further discuss information theoretic bounds for cooperative source coding, present an information theoretic formulation of multi-mode data fusion, and present a framework for study of fundamental limits for network synchronization and position location.

Biography:

Gregory J Pottie is a Professor in the UCLA Electrical Engineering Department and Associate Dean for Research and Physical Resources in the Henry Samueli School of Engineering and Applied Science. He obtained his B.Sc. from Queen's University, Canada, in 1984 and his M.Eng. and Ph.D. from McMaster University, Canada in 1985 and 1988 respectively. From 1989 to 1990 he was employed at Motorola, and since 1991 he has been a faculty member at UCLA. His present research interests include sensor networks and wireless communications systems, and he serves as deputy director of the NSF-funded Center for Embedded Networked Sensing.