Design for Interaction Between Humans and Intelligent Systems During Real-Time Fault Management

Initial results are reported from interdisciplinary projects to provide guidance and assistance for designers of intelligent systems and their human interfaces. The objective is to achieve more effective human-computer interaction (HCI) for real-time fault management support systems. Studies of the development of intelligent fault management systems within NASA have resulted in a new perspective of the user. If the user is viewed as one of the subsystems in a heterogeneous, distributed system, system design becomes the design of a flexible architecture for accomplishing system tasks with both human and computer agents. HCI requirements and design should be distinguished from user interface (displays and controls) requirements and design. Effective HCI design for multi-agent systems requires explicit identification of activities and information that support coordination and communication between agents. We characterize the effects of HCI design on overall system design and identify approaches to addressing HCI requirements in system design. Our results include definition of (1) guidance based on information-level requirements analysis of HCI, (2) high-level requirements for a design methodology that integrates the HCI perspective into system design, and (3) requirements for embedding HCI design tools into intelligent system development environments.