A critical competence for mobile robot systems is the ability to localize itself with respect to the environment. In many real-world applications it is impossible to provide a map of the environment a-priori. Consequently there is a need to equip the robot with facilities for Simultaneous Localisation and Mapping (SLAM) also referred to as Concurrent Mapping and Localisation (CML).

Recently there has been significant progress in methods for SLAM/CML. This summerschool will provide a comprehensive coverage of the key topics and state of the art in SLAM/CML. The summerschool will be a combination of lectures by well known experts in the field and computer exercises (MATLAB, C++/C) involving groups of students.

The summerschool is organised around topics according to:

1. Baysian definition of the problem
   
2. Kalman Filter (KF) interpretation of the problem
   
3. Description of the structure of the problem
   
4. KF implementation of the basic SLAM algorithm
   
5. Bayesian SLAM with and without data association
   
6. Computational issues (KF and Bayesian versions ?)
   
7. Representational issues (points, lines, grids, etc)
   
8. Data association
   
9. Fusion of other data (encoder, inertial, DTED)
   
10. Efficient approximate SLAM algorithms
    
11. SLAM in dynamic environments
    
12. Realisation (case studies)
    
    • Indoor environments with KFs
      
    • Indoor with Monte-carlo methods
      
    • Land vehicles with lasers
      
    • Land vehicles with other sensors (vision, etc)
      
    • Sub-sea implementations
      
    • Air implementations
      
13. Frontiers.
    

The summerschool will involve the following set of experts:

• John Leonard, MIT, Cambridge
  
• Hugh Durrant-Whyte, ACFR, Sydney
• Eduardo Nebot, ACFR, Sydney
• Sebastian Thrun, CMU
• Juan Domingo Tardós Solano, Zaragoza
  
• Wolfram Burgard, Univ of Frieburg
  
• Raja Chatila, LAAS, Toulouse
  
• David Austin, ANU, Canberra
  
• Henrik I Christensen, KTH
• Patric Jensfelt, KTH
  

The summerschool is open to 50 international PhD students. Only registered PhD students will be admitted to the summerschool. It is expected that admitted students have a basic knowledge of linear Kalman Filtering and estimation theory. In addition working knowledge of programming in Matlab and/or C++ is expected. Participation is at a cost of $150, and accommodation will be organised at no extra charge. The participants are required to cover their own travel expenses and local expenses for meals etc.

The summerschool is sponsored by the EU through the European Robotics Research Network -- EURON, the Swedish Foundation for Internationalisation of Research and Education -- STINT and the SSF through its Centre for Autonomous Systems.

Applications for participation in the summerschool must contain information about

1. Name
2. University affiliation
3. Title of project for PhD project
4. Nationality
5. Name of Supervisor
6. Brief summary of current research status (~1/2 page).

Applications must be recieved before June 1. Evaluations of the applications will be completed by June 15.

Applications must be submitted electronically to hic@nada.kth.se

The summerschool is coordinated by an organisation committee:

Generated by Henrik I Christensen, hic@nada.kth.se Sunday, July 7, 2002 22:19