Pamini: A framework for assembling mixed-initiative human-robot interaction from generic interaction patterns

Abstract

Dialog modeling in robotics suffers from lack of generalizability, due to the fact that the dialog is heavily influenced by the tasks the robot is able to perform. We introduce interleaving interaction pat-terns together with a general protocol for task communication which enables us to systematically specify the relationship be-tween dialog structure and task structure. We argue that this approach meets the re-quirements of advanced dialog modeling on robots and at the same time exhibits a better scalability than existing concepts. 1