DICE-R: defining human-robot interaction with composite events

Abstract

Collaborative robots, or cobots, are believed to be a major factor to further increase productivity and to support workers in performing straining, repetitive tasks and tasks that benefit from a third hand. Human workers prefer to interact with cobots in similar ways as with their human peers; using gestures, voice and peripheral awareness. It is thus important to enable cobots to engage in such interactions and to react appropriately. In this paper, we propose a domain-specific textual language, DICE-R, to define human-robot interactions using composite events. DICE-R enables to do this without reference to the automatically generated finite state machines used to recognize these temporal combinations of events from different sources. We introduce the concrete syntax of a DICE-R script, an example application developed using DICE-R, as well as a discussion of how the specified interaction rules are mapped on executable finite state machines.