| |
I am fascinated by distributed computation. Distributed computation
concepts could be applied to the acoustic tracker and others. In addition,
the ideas can expand to mobile computing. Here is reading list of different
ideas.
Filters
-
Condensation
KF goes beyond normal distributions. Possible answer to non-linear
peak detection. Original
paper. Using
the Condensation Algorithm to Implement Tracking for Mobile Robots.
Tracking
Cars in Range Images Using the Condensation Algorithm
The condensation KF is a KF that tracks a spline curve in the prescene
of clutter. The filter uses sampling techniques to track the curve.
-
Unscented KF -- refered
to by Simon Julier ONR. Examples
Uses a dterministic sampling approach. Tech
Report with details.
-
Federated KF
Combines the results of several independently executing KFs. The requirement
is that the state maintained is weakly related or not at all.
-
KF matlab simulation. Non-linear example with acoustic tracker measuring
distance.
-
Wiener filter: offline analog to KF. KF adv. is online computation.
-
Santa barbara perception African Ant path integration
-
federated filters Thesis
-
Gao, Y., Krakiwsky, E.J., Abousalem, M., and McLellan, J.F. (1993). Comparison
and Analysis of Centralized, Decentralized and Federated Filters, Navigation,
Journal of the Institute of Navigation, Vol. 40, No. 1, pp. 69-86. reference
-
Other filters of interest suggested by Mark
Livingston: Covariance
Intersection, recursive total least square estimation.
-
Ulrich Neumann at USC
Distributed estimation
Simon Julier
Design
Issues for an envisioned multi-sensor tracker for tracking computer user’s
pose in an office
|