An Agent Driven Human-centric Interface for Autonomous Mobile Robots

Abstract : One of the challenges in implementing a dynamically autonomous mobile robot is achieving a truly human-centric multimodal interface so that human operators can interact with the robot as naturally as they would with another human. Multiple artificial intelligence techniques may be integrated in a distributed computing system through use of an agent-based architecture. In this effort we utilize an agent-based architecture to achieve a multimodal human-centric interface for controlling a dynamically autonomous mobile robot. Capabilities provided by the architecture include natural language understanding, gesture understanding, localization and mapping for accurate navigation, and use of sensors and maps for spatial reasoning.