Team for Research in
Ubiquitous Secure Technology

Citation
Emily B. Fox, Erik Sudderth, Alan S. Willsky. "Hierarchical Dirichlet Processes for Tracking Maneuvering Targets". Proceedings of the International Conference on Information Fusion, January, 2007.

Abstract
We consider the problem of state estimation for a dynamic system driven by unobserved, correlated inputs. We model these inputs via an uncertain set of temporally correlated dynamic models, where this uncertainty includes the number of modes, their associated statistics, and the rate of mode transitions. The dynamic system is formulated via two interacting graphs: a hidden Markov model (HMM) and a linear-Gaussian state space model. The HMM’s state space indexes system modes, while its outputs are the unobserved inputs to the linear dynamical system. This Markovian structure accounts for temporal persistence of input regimes, but avoids rigid assumptions about their detailed dynamics. Via a hierarchical Dirichlet process (HDP) prior, the complexity of our infinite state space robustly adapts to new observations. We present a learning algorithm and computational results that demonstrate the utility of the HDP for tracking, and show that it efficiently learns typical dynamics from noisy data.

Electronic downloads

• fusion_toappear.pdf · application/pdf · 849 kbytes

• HTML

  Emily B. Fox, Erik Sudderth, Alan S. Willsky. <a
  href="http://www.truststc.org/pubs/260.html"
  >Hierarchical Dirichlet Processes for Tracking
  Maneuvering Targets</a>, Proceedings of the 
  International Conference on Information Fusion, January,
  2007.

• Plain text

  Emily B. Fox, Erik Sudderth, Alan S. Willsky.
  "Hierarchical Dirichlet Processes for Tracking
  Maneuvering Targets". Proceedings of the  International
  Conference on Information Fusion, January, 2007.

• BibTeX

  @inproceedings{FoxSudderthWillsky07_HierarchicalDirichletProcessesForTrackingManeuvering,
      author = {Emily B. Fox and Erik Sudderth and Alan S. Willsky},
      title = {Hierarchical Dirichlet Processes for Tracking
                Maneuvering Targets},
      booktitle = {Proceedings of the  International Conference on
                Information Fusion},
      month = {January},
      year = {2007},
      abstract = {We consider the problem of state estimation for a
                dynamic system driven by unobserved, correlated
                inputs. We model these inputs via an uncertain set
                of temporally correlated dynamic models, where
                this uncertainty includes the number of modes,
                their associated statistics, and the rate of mode
                transitions. The dynamic system is formulated via
                two interacting graphs: a hidden Markov model
                (HMM) and a linear-Gaussian state space model. The
                HMM’s state space indexes system modes, while
                its outputs are the unobserved inputs to the
                linear dynamical system. This Markovian structure
                accounts for temporal persistence of input
                regimes, but avoids rigid assumptions about their
                detailed dynamics. Via a hierarchical Dirichlet
                process (HDP) prior, the complexity of our
                infinite state space robustly adapts to new
                observations. We present a learning algorithm and
                computational results that demonstrate the utility
                of the HDP for tracking, and show that it
                efficiently learns typical dynamics from noisy
                data.},
      URL = {http://www.truststc.org/pubs/260.html}
  }
  

Posted by Emily B. Fox on 15 Jul 2007.
Groups: trust
For additional information, see the Publications FAQ or contact webmaster at www truststc org.