Towards Intelligent Social Robots: Social Cognitive Systems in Smart Environments

Abstract

We propose an institutional robotics approachto the design of socially-aware multi-robot systems, wherecooperation among the robots and their social interactions withhumans are guided using institutions. Inspired by the conceptsstemming from economical sciences, robot institutions serve ascoordination artifacts, which specify behavioral rules that areacceptable or desirable given the situation and which can bereplaced by other rules to enforce new acceptable or desirablebehaviors without changing the robot’s core code. In this paperwe propose a formal methodology for consistent design ofcoordinated multi-robot behaviors intended for use in human-populated environments. We illustrate theoretical concepts withpractical examples. Graph-based formations serve as a basisfor coordinated multi-robot behaviors and concepts from theliterature on human-aware navigation provide social rules thatare enforced by the institutions. Experiments are carried outin a high-fidelity robotic simulator to illustrate the applicationof the theoretical concepts.