LAAS/CNRS                                                        Simon.Lacroix@laas.fr

Robotics and InterationS group                        Tel:  +33 561 33 62 66

7, Avenue du Colonel Roche                             Fax: +33 562 33 64 55

F-31077 Toulouse Cedex 4

France 

1. 
   

2. 
   

3. Environment perception and modeling
   

4. 
   

5. Robot autonomy starts with environment awareness.

   

   Various environment representations, dedicated to specific functionalities (motion or task planning, navigation strategies, motion execution...), must therefore be derived from perceptual data, mainly vision and stereovision in our contributions, acquired either by ground robots or UAVs. 3D geometry naturally plays an essential role in these representations. Recently, we considered the integration of initial data (Geographic Information Systems) in the environment modeling algorithms.
   

6. 
   

7. 
   

8. 
   

9. Vision-based localization and SLAM
   

10. 
    

11. To ensure the spatial consistency of the environment models

    

    data, an as accurate as possible estimation of the robot pose is required: robot localization still remains one of the essential bottleneck for robust autonomous behaviors. Of course, localization and mapping are two intrinsically tied processes, and we have devoted much effort to the resolution of the SLAM problem, using either stereovision, monocular or panoramic vision.
    

12. 
    

13. 
    

14. 
    

15. Autonomous rover navigation
    

16. 
    

17. The achievement of the basic “Goto Goal” navigation

    

    task has been our first objective in the early ‘90s. We always favored an adaptive approach, i.e. an approach in which the rover  selects the best suited motion mode, depending on the environment it has to traverse. We have therefore developed various motion modes, adapted to specific conditions (e.g. flat terrain, rough terrain, path following..), and navigation strategies, that analyze the environment to select and control the appropriate motion mode to apply.
    

18. 
    

19. 

    
    

20. 
    

21. Multi-robot cooperation
    

22. 
    

23. Any robotic mission can be more efficiently androbustly achieved by a team of robots. So far, our work in this area as mainly consisted in conceiving and developing algorithms and architecture for multi-UAVs cooperation and coordination. Our main interest is currently focused on air/ground robotics systems of systems: one of the main challenge is here to build environment representations that integrate aerial and ground data.
    

Research interests: autonomous field robotics