Managing Autonomy Levels in the Ssm/pmad Testbed
Barry R. Ashworth · 2005
Whenever mixing autonomous operations with those of a manual nature, concepts concerning the boundary of operations and responsibility become clouded. The Space Station Module Power Management and Distribution (SSM/PMAD) automation testbed at the NASA, George C. Marshall Space Flight Center in Huntsville, Alabama possesses the need for such mixed-mode capabilities. Martin Marietta, under contract to NASA/MSFC since 1985, is continuing the development of technologies and methodologies for use in the SSM/PMAD automation. Enabling manual seizure of system elements which are currently under automated or autonomous control is a complex problem. Element relations may exist in some systems which do not lend themselves to isolated autonomous management, forcing an unaided manual control user into managing elements beyond the user's capabilities. Also, termination of manual NASA activities within an environment originally intended for autonomous management does not necessarily reinstate the system into its intended autonomous mode. Issues which are surfaced by this manual seizure of autonomously managed elements are: 1) system resource allocation, 2) managing the system component priorities, 3) aiding the user in lessening impacts on any remaining autonomous elements, 4) planning the proper course of events, 5) maintaining an operational system configuration, 6) maintaining any special autonomous capabilities, such as fault recovery, and 7) striving to minimize impacts to any scheduling activities. Here the concept of managing the SSM/PMAD testbed in the presence of changing levels of autonomy is examined. A knowledge-based approach to implementing autonomy management in the distributed SSM/PMAD utilizing a centralized planning system is presented. Its knowledge relations and system-wide interactions are discussed along with the operational nature of the currently functioning SSM/PMAD knowledge-based systems.