You are here

Data Adaptive Fusion for Information Assurance / Information Operations

Award Information
Agency: Department of Defense
Branch: Air Force
Contract: FA9550-10-C-0128
Agency Tracking Number: F09B-T09-0021
Amount: $99,068.00
Phase: Phase I
Program: STTR
Solicitation Topic Code: AF09-BT09
Solicitation Number: 2009.B
Timeline
Solicitation Year: 2009
Award Year: 2010
Award Start Date (Proposal Award Date): 2010-06-15
Award End Date (Contract End Date): 2011-03-14
Small Business Information
8637 East Dunbar Way
Tucson, AZ 85747
United States
DUNS: 962538646
HUBZone Owned: No
Woman Owned: No
Socially and Economically Disadvantaged: No
Principal Investigator
 Harrry Schmitt
 President
 (520) 306-7639
 haschmitt11@gmail.com
Business Contact
 Harrry Schmitt
Title: President
Phone: (520) 306-7639
Email: haschmitt11@gmail.com
Research Institution
 Princeton University
 Robert Calderbank
 
Engineering Quadrangle Olden Street
Princeton, NJ 8544
United States

 (609) 258-3500
 Nonprofit College or University
Abstract

We propose a nine month research program to demonstrate the potential of a set of novel data fusion methods for detecting and localizing distributed patterns of interest in complex networked systems. The technical challenges posed by this problem are significant, but, if they are overcome, the tactical impact is great. Tactical hierarchical sensor networks envisioned or deployed currently operate far from their full potential. The mathematical approaches we propose have the potential to significantly close this performance gap by providing the means to distill information from network data fast, act on it fast and with greater insight than an adversary, and to do that consistently. Our approach is based on recent research by our university subcontractors in two areas: (1) learning the multi-scale dependencies between network node measurements; and (2) constructing sparsifying transforms that facilitates anomaly detection and localization by adaptive aggregation of node measurements. Their method exploits multi-scale spatial structure of the network by performing a hierarchical clustering, and then uses an unbalanced Haar construction to sparsify the network change patterns. Because network activity is summarized by a few large coefficients, the SNR is effectively increased. BENEFIT: The ability to robustly detect and estimate weak distributed patterns in a networked system which are undetectable in the local signatures of individual nodes in the network will enable the Department of Defense and Intelligence Agencies to more fully exploit the inherent potential of complex multi-scale networks in highly dynamic environments. This problem arises in a large number of commercial applications, such as: detecting incipient changes in a spatial field like contamination level monitored by a sensor network, the onset of malicious activity due to a virus spreading in the Internet and identifying statistically significant sets in a gene network or anomalous patterns of social activity. A mixed commercial / government application is the prediction and localization of earthquakes. This is based on the observation that a sudden jump in the number of small quakes in one area seems to be predictive of a larger quake. The challenge is detection this pattern weak pattern against substantial background ‘noise.’

* Information listed above is at the time of submission. *

US Flag An Official Website of the United States Government