Dynamic Structures for Evolving Tactics and Strategies in Team Robotics

Johnson, Jeffrey and Rossi, Ruggero (2018). Dynamic Structures for Evolving Tactics and Strategies in Team Robotics. In: Proceedings of the Conference on Artificial Life and Robotics (ICAROB2018) (Jia, Yingmin; Ito, Takao and Lee, Ju-Jang eds.), ALife Robotics Corp. Ltd., Beppu, Oita, JAPAN, pp. 126–131.

Abstract

The autonomous robot systems of the future will be teams of robots with complementary specialisms. At any instant robot interactions determine relational structures, and sequences of these structures describe the team dynamics as trajectories through space and time. These structures can be represented in algebraic forms that are realizable as dynamic multilevel data structures within individual robots, as the basis of emergent team data structures. Such formalisms are necessary for robots to learn new individual and collective behaviours. The theory is illustrated by the example of robot soccer where robot interactions create structures and trajectories essential to the evolution of new tactics and strategies in a changing environment.

Viewing alternatives

Download history

Item Actions

Export

About

Recommendations