The coordination of multiple autonomous systems using information theoretic political science voting models

In this paper we introduce an approach to solving a coordination problem involving multiple heterogenous autonomous systems operating in a multi-objective mission domain. The distributed coordination mechanism presented applies an information theoretic political science voting model to the dynamic weighted aggregation method, thereby allowing the decision makers (the autonomous systems), to influence the weights assigned to each mission objective. Results from a simulated instantiation of a multiple objective mission are presented, demonstrating the use of the voting mechanism to coordinate actions undertaken by the autonomous systems. The results show a level of overall mission success comparable to that obtained in an 'ideal' centralised mechanism.