[robotics-worldwide] ICRA 2012 Workshop on Stochastic Geometry in SLAM

martin at ing.uchile.cl martin at ing.uchile.cl
Fri Feb 10 19:56:25 PST 2012


Dear all,

We would like to invite submissions for presentation at the ICRA 2012  
workshop on "Stochastic Geometry in SLAM" to be held on 18. May, 2012,  
Saint Paul, Minnesota, USA. Please see the link below for more details:

http://www.ee.uwa.edu.au/~bnvo/icra2012/ICRA_2012_workshop.html


Workshop Aims:

Feature based SLAM is closely related to multi-sensor, multi-target  
filtering. The objective is to jointly estimate a time-varying number  
of targets and their states from sensor measurements with data  
association and detection uncertainty, clutter and noise. Its  
systematic treatment, using random set theory, led to the mathematical  
tools known as Finite Set Statistics (FISST), developed for data  
fusion and estimation of random sets.

Random vector based SLAM is known to be extremely fragile in the  
presence of feature detection and data association uncertainty.  
Therefore recent research which has applied the FISST framework to  
autonomous map representations, has shown that under this paradigm,  
the fragile problems of map management and feature to feature  
association can be eliminated.  This framework will be the subject of  
this workshop. Workshop topics will include global localisation,  
Rao-Blackwellised and multi-robot SLAM which jointly consider false  
alarms, missed detections and spatial sensor uncertainty. Experimental  
results in challenging outdoor and marine environments will be  
demonstrated.

Topics:

1. An introduction to Finite Set Statistics (FISST). Representing  
uncertainty with Random Finite Sets (RFSs) versus random vectors.
2. Random set theoretic methods in global localisation.
3. Rao-Blackwellised RFS Bayesian SLAM, demonstrating experiments in  
cluttered outdoor and marine environments.
4. A Single-Cluster Multi-Object Approach to Simultaneous Localisation  
and Mapping.
5. Extending the Bayesian RFS framework to multi-robot SLAM.
6. Robust multi-target filtering and SLAM.
7. Performance error metrics for SLAM - Evaluation map estimation  
error in its entirety.
8. Map building using FISST.

Invited Speakers:

Ronald P. Mahler, Lockheed Martin Tactical Systems, Minnesota, USA.
Daniel E. Clark, Herriot Watt University, UK.
Adrian Bishop, ANU/NICTA, Australia.
Diluka Moratuwage, NTU, Singapore.
Umut Orguner, Link?ping University, Sweden
Ba-Ngu Vo, University of Western Australia.
Martin Adams, Universidad de Chile, Santiago, Chile.

Important Dates:

Paper Submission:            12. March 2012

Notification of Acceptance:  16. April 2012

Final Paper Submission:      1. May 2012

Workshop Date:               18. May 2012


Best regards,


Workshop organisers:

Martin Adams (University of Chile): martin at ing.uchile.cl

Ba-Ngu Vo (University of Western Australia): ba-ngu.vo at uwa.edu.au

------------------------------------------
Dr. Martin Adams
Professor of Electrical Engineering
Universidad de Chile
Dept. Electrical Engineering
Av. Tupper 2007, 837?0451 Santiago, Chile

Phone: +56?2?978?4209
Fax: +56?2?672?0162
Email: martin at ing.uchile.cl
url: http://www.cec.uchile.cl/~martin
------------------------------------------



More information about the robotics-worldwide mailing list