Date: February 16, 2007

Time: 3:00 p.m.

Location: Centergy One 5186

Speaker(s): Milind Borkar

Title: A Distributed Monte Carlo Method for Initializing State Vector Distributions in Heterogeneous Smart Sensor Networks

Abstract: The objective of this talk is to demonstrate how an underlying system's state vector distribution can be determined in a distributed heterogeneous sensor network with reduced subspace observability at the individual nodes. We show how the network, as a whole, can be made capable of observing the target state vector even if the individual nodes are not capable of observing it locally. The algorithm presented can generate the initial state vector distribution for networks with a variety of sensor types as long as the measurements at the individual nodes are functions of the target state vector. Initialization is accomplished through a novel distributed implementation of the particle filter that involves serial particle proposal and weighting strategies that can be accomplished without sharing raw data between individual nodes in the network. If multiple events of interest occur, their individual states can be initialized simultaneously without requiring explicit data association across nodes. The resulting distributions can be used to initialize a variety of distributed joint tracking algorithms.

Bio: Milind Borkar received the B.S. and M.S. degrees in Electrical Engineering from the Georgia Institute of Technology in 2002 and 2005. He is currently a Ph.D. student under the advisement of Dr. James McClellan. His research interests include Markov chain Monte-Carlo techniques, target tracking in a Bayesian framework, distributed smart sensor networks, distributed data fusion and the use of multimedia in education.

Date:	February 16, 2007
Time:	3:00 p.m.
Location:	Centergy One 5186
Speaker(s):	Milind Borkar
Title:	A Distributed Monte Carlo Method for Initializing State Vector Distributions in Heterogeneous Smart Sensor Networks
Abstract:	The objective of this talk is to demonstrate how an underlying system's state vector distribution can be determined in a distributed heterogeneous sensor network with reduced subspace observability at the individual nodes. We show how the network, as a whole, can be made capable of observing the target state vector even if the individual nodes are not capable of observing it locally. The algorithm presented can generate the initial state vector distribution for networks with a variety of sensor types as long as the measurements at the individual nodes are functions of the target state vector. Initialization is accomplished through a novel distributed implementation of the particle filter that involves serial particle proposal and weighting strategies that can be accomplished without sharing raw data between individual nodes in the network. If multiple events of interest occur, their individual states can be initialized simultaneously without requiring explicit data association across nodes. The resulting distributions can be used to initialize a variety of distributed joint tracking algorithms.
Bio:	Milind Borkar received the B.S. and M.S. degrees in Electrical Engineering from the Georgia Institute of Technology in 2002 and 2005. He is currently a Ph.D. student under the advisement of Dr. James McClellan. His research interests include Markov chain Monte-Carlo techniques, target tracking in a Bayesian framework, distributed smart sensor networks, distributed data fusion and the use of multimedia in education.