Autonomous Construction by a Mobile Robot in Unknown Environments with Scarce Resources

• Presented at: 4th International Conference on Cognitive Systems, ETH Zurich, Switzerland, January 27 & 28, 2010
• Publication date: 2010

Autonomous construction by mobile robots would be useful in various situations, such as in outer space, in hazardous environments, but also for the building industry. Current works tackle simplified scenarios where environment is flat and resources readily available; moreover robots build simple structures. However, target applications would feature complex 3D environments, remote resources, and would require the construction of multi-layer structures of various types. In this poster, we show our current work in pursuing a step toward such applications. We propose a simple experimental setup where a miniature mobile robot senses its environment and finds the right course of action to build structures. Resources are remote, requiring the robot to first locate them, then to create a way to fetch them and finally to build the structure. The robot performs SLAM to build a map and uses an HTN planner to choose its actions. It manipulates its environment through a magnetic gripper.

Note: This version correspond to the printer version presented at the conference. The version in the proceedings is slightly older and does not present the latest results.

Keywords: [MOBOTS] ; autonomous construction ; mobile robot ; unknown environments ; scarce resources ; marxbot ; aseba ; cognition ; morphological operation ; image processing ; vision ; manipulation ; HTN planning ; action ; aseba ; state machines

Reference

• LSRO-POSTER-2010-001

• URL: http://cogsys2010.ethz.ch/

Record created on 2010-01-21, modified on 2012-03-21