Conversational human-swarm interaction using IBM Cloud

Abstract

Swarm robotics is an approach to the coordination of large numbers of robots that has become an increasingly popular field of research in recent years, not least because properly engineered robot swarms are scalable, flexible, and robust, making them an attractive alternative to single-robot systems in many application domains. Since its inception, the field of swarm robotics has grown beyond its roots in purely decentralised control inspired by social insect behaviour, now often utilising hybrid centralised/decentralised control architectures that incorporate human operators who guide swarm actions during tasks such as firefighting, or the localisation of radiation sources. This kind of human-swarm interaction has attracted significant interest from the research community, spawning an entire sub-field of its own that investigates how human operators, supervisors, and team-mates can interact with robot swarms and receive feedback from them. To date, human-swarm control methods such as the use of graphical user interfaces and spatial gestures have received much attention, but there has been little investigation into the potential of controlling swarm robotic systems with an operator’s voice. The few studies that have explored this idea are restricted to the use of specific predefined phrases that the human operator is required to learn, resulting in interactions that are unnatural in comparison to the way a human would normally express themselves in speech. In this paper, we present a novel architecture for conversational human-swarm interaction that addresses these issues, allowing swarm robotic systems to be engineered in such a way that a human operator can guide a swarm using spoken dialogue in a more natural manner.