Objective:

NASA must increasingly rely on distributed teams of experts serving multiple projects. Supporting their activity is challenging, particularly during off-nominal situations with rapidly changing status, tasks, roles, locations, and preferences. Agent-based systems can help with situational awareness, targeted information dissemination, and coordination. This research task will develop an information architecture to support nominal and off-nominal operations by small, flexible, loosely-coupled teams of cooperating humans and software agents in dynamic mission environments. The architecture will be demonstrated for simulated operation of a life support system. Assistant and proxy agents will support teamwork and integrate with management agents that handle routine operations. Information and control will be distributed across multiple agents, with synchronous and asynchronous communication. Interactions could be steered by either a user or an agent, with the system verifying the safety and appropriateness of directed actions. In cases of conflict, the system will balance user needs and priorities against agent goals in deciding how to act. An interface will allow display of system events and performance, plus selected knowledge, plans, and actions of agents. The architecture will also be used by its own developers, as an example of agent-based support for software development.

Applications:

Agent-based support tools for anomaly response teams in dynamic situations, and for distributed routine operations; tested for life support system scenarios.

NASA Benefit:

As missions increase in number, complexity, and duration, NASA must increasingly rely on teams of distributed experts who divide their time among multiple projects. The proposed architecture will reduce the mission impact of off-nominal situations, by means of better situational awareness (via prompt and appropriate information dissemination and coordination). The results should be useful throughout NASA mission operations. The integrated demonstration will incorporate HCC-funded projects in mission control, life support, and the Space Medical Agents Resource Technology (SMART) console for flight surgeons and biomedical engineers, with opportunities for technology insertion from other projects. The software developers will be using the coordination architecture for their own work, ensuring rapid development and thorough testing of the system. Investigators will document the operations concept and architecture, as well as the process and products of human-centered analysis that were used, to provide guidance for other software development efforts.

Keywords:

mission operations, distributed team coordination, software agents, mixed-initiative control

Images:

PI slides.

Prior Technology:

Communications dependent on participant awareness of staff assignments, locations, availability, task status, etc.

FY04 Milestone:

Coordination infrastructure and robust interface designs.

FY04 Quadchart Slide:

HCC_DIR_Malin_IntOps.ppt.

Accomplishments:

Testbed infrastructure; intelligent assistant prototypes; teamwork tool suite; Intelligent System Management Agents (ISMA) to support distributed teams; human-centered methods for prototyping and evaluating team ops concepts; Intelligent Briefing and Response Assistants (IBRA) to manage information; workspace and issue-tracking software tool for Space Shuttle Mission Control; anomaly handling study; infusion of teamwork tools; demonstrations with simulated systems.

Contacts:

Jane T. Malin (PI), JSC Automation, Robotics, and Simulation.
Hongbin Wang (Co-I), U. Texas Health Science Center at Houston.
Debra K. Schreckenghost (Co-I), Metrica at Johnson Space Center.
David M. Kortenkamp (Co-I), Metrica at Johnson Space Center.
Jiajie Zhang (Co-I), U. Texas Health Science Center at Houston.
Jack W. Smith (Co-I), UT Health Science Center and JSC.
Kathy A. Johnson-Throop (Co-I), UT Health Science Center and JSC.