SuperLectures.com

PARTICLE FLOW FOR NONLINEAR FILTERS

Particle Filtering for High Dimensional Problems

Full Paper at IEEE Xplore

Presented by: Fred Daum, Author(s): Fred Daum, Jim Huang, Raytheon, United States

We solve the well known problem of particle degeneracy using a new theory that computes Bayes' rule using particle flow rather than a pointwise multiplication of two functions. Our new filter does not resample particles, and it does not use any proposal density. The computational complexity is four orders of magnitude faster than standard particle filters, because we never resample particles. The filter accuracy is typically several orders of magnitude better than the extended Kalman filter (EKF) or UKF for difficult nonlinear problems. The particle flow is designed by solving a linear first order highly underdetermined PDE. We discuss a dozen distinct solutions for this PDE, including four completely new solutions.


  Speech Transcript

|

  Slides

Enlarge the slide | Show all slides in a pop-up window

0:00:16

  1. slide

0:00:25

  2. slide

0:00:59

  3. slide

0:01:55

  4. slide

0:02:20

     3. slide

0:02:27

     4. slide

0:02:34

  5. slide

0:03:07

  6. slide

0:03:28

  7. slide

0:05:34

  8. slide

0:06:15

  9. slide

0:06:58

 10. slide

0:07:59

 11. slide

0:08:58

    11. slide

0:09:54

 12. slide

0:10:51

 13. slide

0:11:23

 14. slide

0:12:19

 15. slide

0:13:50

 16. slide

0:14:00

 17. slide

0:15:00

 18. slide

  Comments

Please sign in to post your comment!

  Lecture Information

Recorded: 2011-05-26 16:15 - 16:35, Club D
Added: 18. 6. 2011 23:52
Number of views: 45
Video resolution: 1024x576 px, 512x288 px
Video length: 0:21:50
Audio track: MP3 [7.39 MB], 0:21:50