As robots are gradually leaving highly structured factory environments and moving into human populated environments, they need to possess more complex cognitive abilities. Not only do they have to operate efficiently and safely in natural populated environments, but also be able to achieve higher levels of cooperation and interaction with humans. The Autonomous City Explorer (ACE) project envisions to create a robot that will autonomously navigate in an unstructured urban environment and find its way through interaction with humans. To achieve this, research results from the fields of autonomous navigation, path planning, environment modeling, and human-robot interaction are combined.
ACE has its own channel on YouTube (www.youtube.com/user/AutonomousCityExplor), where there are videos about the project and experiment, a trailer, and the acceptance speech for the IJCAI-09 AI Video Competition "most innovative video" award.
Navigation and localization
- Particle filter based SLAM and moving objects tracking
- Bayesian framework for behavior selection under uncertainty in dynamic environments
- Human-robot communication user-studies
- Communication of map knowledge
- Gesture based communication of navigational knowledge
Object classification/segmentation in 3D laser data
- Efficient surface normal estimation for 3D point clouds obtained by a mobile robot
- Segmentation of 3D laser data point clouds with graph-theoretic and template-based approach
The evolution of ACE.
July 2008: ACE went autonomously through the Munich
pedestrian area to the Marienplatz.
August 2008: ACE went autonomously from the lab to the Marienplatz in Munich. It did not use any map knowledge or GPS and found its way solely from information given from passers-by. (Everybody who gave directions to ACE: thank you.)