A multi-agent based architecture for an adaptive human-robot interface

The increasing capability of robotic technology has resulted in widespread interest in research on mobile robots, especially with regard to both teleoperation and autonomous operation. However, responses of many autonomous robotic systems to unanticipated circumstances are typically not robust enough to be used in the field. The skills of the human user could fill this gap. Across a wide range of real-world tasks, efficient teams of humans and robots may provide a valuable alternative for task accomplishment. A cooperative system of humans and robots are expected to be more capable than a team of autonomous robots in achieving real-world applications. To achieve this human-robot teaming, robust user interfaces are, therefore, of fundamental importance in order for useful communication to be exchanged in an effective way between the human and robot on a variety of levels.; While past work on mobile robots has predominately focused on issues such as navigation and sensing, relatively few robotic systems are equipped with robust user interfaces that have the adaptability to adjust interfaces based on the user intension (e.g. mission goals), operation requirements, and the current robot situations. This dissertation developed a new multi-agent based architecture for an adaptive human-robot interface. This dissertation differs from previous research in human-robot supervisory systems in four important ways. First, the system architecture is based on a multi-agent system that should permit porting to any mobile robot platform with a variety of sensors with relative ease. Second, this research developed a new Spreading Activation Network (SAN) Generator that could automatically create the robot's behavior selection network according to the mission command and the robot capabilities. Third, the distributed, agent-based user interface provides an “Event Triggered Adaptation” where one agent generates an event based upon the context of operation to initiate the change in the interface. Fourth, the architecture allows robot-to-human communication whereby the robot is able to could query the user for assistance when needed.