CRV 2011 SLAM Camp
A free 'SLAM camp' was held on at CRV 2010 (in Ottawa).
A similar camp may be held with CRV 2011, depending on availability of participants.
Simultaneous Localization and Mapping (SLAM) is a probabilistic technique that attempts to
accurately localize a vehicle (without the aid of GPS) while generating a map of its environment.
Unique environmental features (landmarks) are extracted from sensor data (laser scanner, LIDAR,
video CCD, for example) and are tracked across time and used as constraints to improve a robot's pose
estimate. This improved pose estimate is, in turn, used to improve the estimated pose/position of the
unique landmarks. SLAM has been actively investigated in the robotics industry for the last 15 years.
Description
A full day workshop exploring Simultaneous Localization and Mapping (SLAM) may be held in
conjunction with the Eighth Canadian Conference on Robot Vision (CRV 2011). The workshop will
explore concepts related to SLAM including basic formulation through to current state of the art
techniques. Math is to be kept to a minimum with the focus on practical examples and key concepts.
Goals of the workshop are to familiarize researchers with the core concepts of SLAM, to generate
interest in academia for the SLAM problem, and to bring together Canadian SLAM researchers to
explore possible collaborations.
We are inviting your attendance and participation in the SLAM CRV workshop either as an attendee or
speaker. If you wish to present in one of the areas listed below or have suggestions for other SLAM
related topics, please contact Jack Collier at jack.collier@drdc-rddc.gc.ca.
CRV SLAM Workshop Topics
Probability Theory
Basic SLAM formulation (EKF)
Feature Detection
Advanced Data Association Techniques
Particle Filter SLAM (FastSLAM)
Loop Closing techniques
Sub-mapping/hybrid mapping techniques
Vision Based SLAM
Appearance Only SLAM
Cooperative SLAM
SLAM in the field (Practical Considerations)
Future Research Directions
|