SWARMORPH-script: A Language for Arbitrary Morphology Generation in Self-Assembling Robots


[ Overview ] [ Photos ] [ Videos ] [ Publications ] [ Contact ]


  Overview


In a multi-robot system the physical limitations of the individual robots can be overcome when robots have the ability to self-assemble, thus forming larger composite robotic entities. However, existing robotic systems capable of self-assembly have little or no control over the morphology of the self-assembled entities. This restricts the adaptability of such systems, since robots can carry out certain tasks more efficiently if their morphology is specialized to the task.

In this study, we extend the distributed mechanism presented in Christensen et al. 2007 that allows autonomous mobile robots to self-assemble into specific morphologies. The morphologies are generated based on a simple language, SWARMORPH-script, that allows for concise descriptions of the rules that govern the distributed morphology growth process. Local visual communication allows physically connected robots to send and receive strings. A string can be a rule identifier that triggers execution of predefined logic for extending a morphology. Alternatively, whole scripts can be communicated and subsequently executed on the receiving robot. On real self-propelled robots capable of self-assembly, we demonstrate how specific morphologies can be constructed, how the size of a morphology can be regulated, and how multiple identical morphologies can be assembled. We also show how the transmission of entire scripts gives the robots the capacity to participate in the formation of morphologies of which they had no apriori knowledge.

    

    




  Photos


              


              


              






  Videos





Communication test.
MPEG1 ~43MB

Horseshoe morphology.
MPEG1 ~22MB



Shovel morphology.
MPEG1 ~33MB

Square morphology.
MPEG1 ~53MB



Minisquare morphologies.
MPEG1 ~33MB

Lines-of-three morphologies.
MPEG1 ~32MB



Communicated-line replicating
morphology. MPEG1 ~43MB

Communicated-step replicating
morphology. MPEG1 ~47MB



Communicated-triangle replicating
morphology. MPEG1 ~36MB




  Publications


A paper on this study has been submitted to the Swarm Intelligence journal.




  Contact


Webpages:

Swarm-bots: www.swarm-bots.org
Swarmanoid: www.swarmanoid.org
Anders' homepage: http://iridia.ulb.ac.be/~alyhne
Rehan's homepage: http://iridia.ulb.ac.be/~rogrady
Marco's homepage: http://iridia.ulb.ac.be/~mdorigo

Address:

IRIDIA - ULB
50 Avenue F. Roosevelt - CP 194/9
1050 Bruxelles
Belgium



This is the end - now go program a robot or a hundred...