Linguistic encoding of temporal relations in robotic system scenarios

This paper introduces an approach to linguistic encoding of temporal relations in robotic system scenarios. The underlying idea is that a dialogue system that manages the interaction between the user and the external robotic system should be able (i) to understand and adopt a scenario defined in natural language, and (ii) to express a scenario, stored in a computationally appropriate form, in natural language. The paper proposes a language-independent notion of temporal structure, reports a corpus-based study on the role of temporal prepositions in signaling temporal relations in the Serbian language, and discusses the implementation of the prototype dialogue system.