SLAM SUMMER SCHOOL 2009

The 4th Summer School in Simultaneous Localisation and Mapping (SLAM) was hosted by the Australian Centre for Field Robotics, University of Sydney on 20-23 January 2009.

Background

For more than a decade the robotics community has focussed extensively on the concept of Simultaneous Localisation and Mapping (SLAM) and it has become one of the central themes in the implementation of many autonomous systems. SLAM is a set of algorithms, techniques, heuristics and paradigms which come together to solve a very difficult problem - “how can a mobile robot navigate in an a-priori unknown environment”.

There is a plethora of research activity in this area, most of which is focussed on indoor applications, with simple robots and simple sensors, operating in structured environments. Some of the techniques developed have worked their way into more complex mobile robot systems operating in outdoor, natural and semi-structured environments, however with limited success. Despite the widespread use of GPS in such environments it has its limitations and many systems require a localisation solution that can operate where GPS is denied, cannot be accessed or with limited availability. Classic examples include planetary exploration, underwater robotics, systems operating under extensive canopy or around tall buildings, where GPS has been intentionally jammed, and in disaster zones. Even when GPS is available, SLAM provides a means of incorporating redundancy in the navigation solution for the purposes of reliability.

Objectives

The goal of the summer school was to provide lectures and laboratory work on the basics as well as the state-of-the-art of SLAM, with a bent towards outdoor applications and perception in outdoor environments. A number of leading international researchers presented their ideas and solutions to this difficult problem, focussing on ideas of machine learning for perception and situational awareness, large scale SLAM and Active-SLAM. The summer school provided the opportunity for research students who were new to, or interested in, this exciting area to see what methods are being employed and what systems have been demonstrated, as well as to get hands on experience with some of the algorithmic methods. It also provided the opportunity to interact with fellow researchers in this field with the aim of generating collaborative initiatives.

Click on the links below to get a copy of the presentations.

Presentations

Tuesday 20 January 2009

09:00-10:00 Simultaneous Localisation and Mapping: Probabilistic Formulation - Tim Bailey
Tim Bailey Slides
Tim Bailey Presentation Recording 185MB

10:00-11:00 From Bayes' rule to the EKF SLAM - Jose Neira
Jose Neira Slides
Jose Neira Presentation Recording 170MB

11:15-12:15 Advanced EKF-SLAM: Building Large Maps - Juan Tardos
Juan Tardos Slides
Juan Tardos Presentation Recording 177MB

12:15-13:15 Active and Action Based SLAM - Mitch Bryson
Mitch Bryson Slides
Mitch Bryson Presentation Recording 177MB


Wednesday 21 January 2009

09:00-11:00 3D Graphical Models for SLAM - Frank Dellaert
Frank Dellaert Slides
Frank Dellaert Presentation Recording 1 168MB
Frank Dellaert Presentation Recording 2 151MB

11:15-12:15 Why Perception is Hard - Steve Scheding
Steve Scheding Slides
Steve Scheding Presentation Recording 141MB

12:15-13:15 Techniques for 3D Mapping Theory - Wolfram Burgard
Wolfram Burgard Slides
Wolfram Burgard Presentation Recording 192MB


Thursday 22 January 2009

09:00-10:00 Future Challenges for Robotic Perception - Luke Fletcher
Luke Fletcher Slides
Luke Fletcher Presentation Recording 150MB

10:00-11:00 Fabio Ramos
Fabio Ramos Slides
Fabio Ramos Presentation Recording 164MB

11:15-12:15 Fabio Ramos
Fabio Ramos Slides
Fabio Ramos Presentation Recording 158MB

12:15-13:15 Appearance-Based Methods for Loop Closing - Mark Cummins
Mark Cummins Slides
Mark Cummins Presentation Recording 120MB


Friday 23 January 2009

09:00-10:00 Techniques for 3D Mapping Applications - Wolfram Burgard
Wolfram Burgard Slides
Wolfram Burgard Presentation Recording 171MB

10:00-11:00 UGV & AUV SLAM and Mapping - Simon Lacroix
Simon Lacroix Slides
Simon Lacroix Presentation Recording 177MB

11:15-12:15 Urban Mapping - Luke Fletcher
Luke Fletcher Slides
Luke Fletcher Presentation Recording 147MB

12:15-13:15 AUV SLAM - Ian Mahon
Ian Mahon Slides
Ian Mahon Presentation Recording 148MB

14:00-15:00 SLAM on an Airborne Vehicle - Mitch Bryson
Mitch Bryson Slides
Mitch Bryson Presentation Recording 137MB

15:00-16:00 CORD Ground Vehicle Programme - Steve Scheding
Steve Scheding Slides
Steve Scheding Presentation Recording 168MB