Uninformative Prior Multiple Target Tracking Using Evidential Particle Filters
Author(s)
Worthy, Johnny L., III
Holzinger, Marcus J.
Advisor(s)
Editor(s)
Collections
Supplementary to:
Permanent Link
Abstract
Space situational awareness requires the ability to initialize state estimation from
short measurements and the reliable association of observations to support the
characterization of the space environment. The electro-optical systems used to observe
space objects cannot fully characterize the state of an object given a short,
unobservable sequence of measurements. Further, it is difficult to associate these
short-arc measurements if many such measurements are generated through the
observation of a cluster of satellites, debris from a satellite break-up, or from spurious
detections of an object. An optimization based, probabilistic short-arc observation
association approach coupled with a Dempster-Shafer based evidential
particle filter in a multiple target tracking framework is developed and proposed
to address these problems. The optimization based approach is shown in literature
to be computationally efficient and can produce probabilities of association, state
estimates, and covariances while accounting for systemic errors. Rigorous application
of Dempster-Shafer theory is shown to be effective at enabling ignorance to
be properly accounted for in estimation by augmenting probability with belief and
plausibility. The proposed multiple hypothesis framework will use a non-exclusive
hypothesis formulation of Dempster-Shafer theory to assign belief mass to candidate
association pairs and generate tracks based on the belief to plausibility ratio.
The proposed algorithm is demonstrated using simulated observations of a GEO
satellite breakup scenario.
Sponsor
Date
2017-09
Extent
Resource Type
Text
Resource Subtype
Paper
Rights Statement
Unless otherwise noted, all materials are protected under U.S. Copyright Law and all rights are reserved