Multirobot Systems and Physical Data Structures
Papers from the AAAI 2011 Spring Symposium
Dylan Shell and James McLurkin, Program Cochairs
Multirobot systems can represent and manipulate physical representations of information by modifying their environment or changing their positions. Information necessary for agent coordination or collective task performance can be externalized into the environment, and then used by other robots when and where necessary. Systems are able to exploit the locality of this external information, its persistence, its dynamics, or how it affects the task mechanics to complete tasks with higher efficiency or enable radically different solutions. However, there is no unified theory of the cost and complexity of communicating or storing information in this way.
This symposium brought researchers together who study multiagent and multirobot systems and are interested in unconventional methods for exploiting state information, the application of existing approaches in order to understand externalized memory, exploitation of implicit communications between agents, as well as broader efforts which reduce sensing, communication, or computational requirements. Contributions presented include theoretical discussion of principles for such work, and practice of these techniques.