Interpreting Suspicious and Coordinated Behavior in a Sensor Field
Loading...
Authors
Rowe, Neil C.
Subjects
Advisors
Date of Issue
2008-08
Date
August 2008
Publisher
Monterey, California. Naval Postgraduate School
Language
Abstract
We report on recent work we have done on detection of two kinds of militarily
interesting behavior in an urban battlespace, detection of suspicious behavior and
detection and classification of coordinated movements of groups of people. The first
is important in detecting terrorism and IED emplacement, and the second is
important in detecting military adversaries and what they are doing. Our approaches
use only "pose" information, the locations and orientations of people within the
sensor field, as extracted from tracking by a fusion of various nonimaging sensing
modalities. Restriction to nonimaging sensors saves money, and restriction to pose
information avoids most of the serious privacy concerns. We first explain our
approach to tracking using signal strengths alone. From experiments with both
staged and nonstaged behavior in a public area, we found that the most useful clue to
suspicious behavior was the norm of the acceleration vector averaged over several
different time scales. With detection and classification of groups of people, by
contrast, no single metric was as good as combinations of metrics. We are exploring
a variety including average distances between people, uniformity of distances,
linearity of the positions of people, number of clusters of people, number of
directions in which they can see, overall visibility, average speed of the group, and
uniformity of the speed of the group. A key challenge is to make these metrics
scale-free as with the acceleration vector analysis.
Type
Conference Paper
Description
This paper appeared in the Proceedings of the Meeting of the Military Sensing Symposium Specialty Group on
Battlespace Acoustic and Seismic Sensing, Magnetic and Electric Field Sensors, Laurel, MD, August 2008.
Series/Report No
Department
Organization
Identifiers
NPS Report Number
Sponsors
supported in part by the National Research Council under their Research Associateship Program at the Army Research Laboratory, in part by the National Science Foundation under the EXP Program, and in part by the BASE-IT Project sponsored by the Office of Naval Research
Funding
Format
Citation
Distribution Statement
Approved for public release; distribution is unlimited.