Reimagining Trustworthy Robot Fleets with Animal Analogies

Abstract

In the future, multi-agent robot fleets will be important for domains like agriculture, space exploration, and air combat. Trust of human-machine teams is needed to make the teams resilient to the faults of both human and robot teammates. Trust in multi-agent systems is often fragile: if any agent in the system is less reliable than the others, people will stop interacting with all of them. Studying relationships in human-animal systems can provide useful insights into designing humanrobot systems. We present a method for gathering insight into how humans, working with animal systems think about the relationships between the individuals and the whole, and suggest how animal system models can be used as analogies and practical design features for the design of robot systems in order to increase trust. Using a more-than-human approach in design research phase of human-robot interaction, supports more secure collaboration between humans and robot systems.