Integrating Intelligent Systems
into Human Teams

For this MURI project we have conducted multidisciplinary research aimed at incorporating intelligent assistants into human teams to increase the effectiveness of team decision making in joint planning tasks. The research paradigm of coordination based on developing and maintaining shared mental models provides a common perspective linking the intelligent agents research at Carnegie Mellon University, the research in multi-media information delivery of the Software Engineering Institute, the team performance studies at the Naval Air Warfare Training Systems Division, and the task directed orientation to cognitive modeling and information presentation of the University of Pittsburgh and the NRL.

To aid in the fast paced, multi user environment of joint mission planning, assistants develop capacities for anticipating the information needs of their human collaborators, preparing and communicating task information, adapting to changes in situation and capabilities of other team members, and effectively supporting team member mobility.

In developing adaptive and self organizing collections of Intelligent Agents who possess models for discriminating and communicating situational distinctions salient to humans and to a team's mission, we have addressed the issues in incorporating intelligent assistants into human teams. We have developed agent and system architectural principles and reusable intelligent agent software components enabling (1) seamless integration of information access with user-centered problem solving and decision support, (2) active monitoring and intelligent caching of environmental information so that users or user-delegated intelligent assistants can acquire up-to-date situation evaluation and therefore increase the relevance and quality of decision making, and (3) adaptive formation of working teams among intelligent assistants ``on-demand'' depending on the information requirements of a particular task.

Our research (1) provides a methodology for cognitive modeling of member and team tasks, (2) provides principles for developing and supporting high performance teams, (3) contributes to the understanding of agent roles and human agent interactions in teams composed of humans and intelligent agents, (4) produces a set of architectural principles and techniques that allow agents to cooperatively inter-operate to provide integrated and adaptable information access, integration and decision support, (5) results in a set of reusable agent structuring components that will allow system developers to rapidly construct new agents for different end-user applications, and (6) provides techniques and tools for creating, visually representing and accessing large scale multimedia, multimodal information including structured data as well as speech, text, images and video.

Our findings have implications for other types of planning teams that are comprised of multi-disciplinary experts. Examples include civilian emergency response, management, and single service military teams.

Multidisciplinary University Research Initiative (MURI)
Principal Investigator: Katia Sycara
Sponsored by: Office of Naval Research (ONR)
ONR Contact: Michael Shneier
© 1998 Carnegie Mellon Robotics Institute


Carnegie Mellon School of Computer Science Carnegie Mellon Robotics Institute DefenseLINK Office of Naval Research MURI Welcome University of Pittsburgh School of Information Sciences