Bayesian filtering over compressed appearance states

Douillard, Bertrand, Upcroft, Ben, Kaupp, Tobias, Ramos, Fabio and Durrant-Whyte, Hugh (2007). Bayesian filtering over compressed appearance states. In: Matthew Dunbabin and Mandyam Srinivasan, Proceedings of the 2007 Australasian Conference on Robotics & Automation. Australasian Conference on Robotics and Automation 2007, Brisbane, Australia, (). 10 - 12 December 2007.


Author Douillard, Bertrand
Upcroft, Ben
Kaupp, Tobias
Ramos, Fabio
Durrant-Whyte, Hugh
Title of paper Bayesian filtering over compressed appearance states
Conference name Australasian Conference on Robotics and Automation 2007
Conference location Brisbane, Australia
Conference dates 10 - 12 December 2007
Proceedings title Proceedings of the 2007 Australasian Conference on Robotics & Automation
Publication Year 2007
Sub-type Fully published paper
ISBN 978-0-9587583-9-0
Editor Matthew Dunbabin
Mandyam Srinivasan
Total pages 10
Language eng
Abstract/Summary This paper presents a framework for perform- ing real-time recursive estimation of landmarks’ visual appearance. Imaging data in its origi- nal high dimensional space is probabilistically mapped to a compressed low dimensional space through the definition of likelihood functions. The likelihoods are subsequently fused with prior information using a Bayesian update. This process produces a probabilistic estimate of the low dimensional representation of the landmark visual appearance. The overall filter- ing provides information complementary to the conventional position estimates which is used to enhance data association. In addition to robotics observations, the filter integrates human observations in the appear- ance estimates. The appearance tracks as com- puted by the filter allow landmark classifica- tion. The set of labels involved in the clas- sification task is thought of as an observation space where human observations are made by selecting a label. The low dimensional appearance estimates re- turned by the filter allow for low cost com- munication in low bandwidth sensor networks. Deployment of the filter in such a network is demonstrated in an outdoor mapping applica- tion involving a human operator, a ground and an air vehicle.
Subjects 1109 Neurosciences
09 Engineering
Q-Index Code EX

 
Versions
Version Filter Type
Citation counts: Google Scholar Search Google Scholar
Access Statistics: 26 Abstract Views  -  Detailed Statistics
Created: Wed, 14 Apr 2010, 14:37:37 EST by Laura McTaggart on behalf of Faculty Of Engineering, Architecture & Info Tech