Magnusson, Martin Doherty, Patrick 2009 (English)In: The Swedish AI Society Workshop 2009, SAIS 2009 / [ed] Fredrik Heintz and Jonas Kvarnström, Linköping University Electronic Press, Linköpings universitet, 2009, 39-48Conference paper (Other academic)
Cooperation is a complex task that necessarily involves communication and reasoning about others’ intentions and beliefs. Multi-agent communication languages aid designers of cooperating robots through standardized speech acts, sometimes including a formal semantics. But a more direct approach would be to have the robots plan both regular and communicative actions themselves. We show how two robots with heterogeneous capabilities can autonomously decide to cooperate when faced with a task that would otherwise be impossible. Request and inform speech acts are formulated in the same first-order logic of action and change as is used for regular actions. This is made possible by treating the contents of communicative actions as quoted formulas of the same language. The robot agents then use a natural deduction theorem prover to generate cooperative plans for an example scenario by reasoning directly with the axioms of the theory.
Linköping Electronic Conference Proceedings, ISSN 1650-3686 (print), 1650-3740 (online) ; 35
National CategoryEngineering and Technology
Identifiersurn:nbn:se:liu:diva-28259 (URN)oai:DiVA.org:liu-28259 (OAI)diva2:249049 (DiVA)
25th annual workshop of the Swedish Artificial Intelligence Society (SAIS 2009), 27-28 May 2009, Linköping, Sweden