Intelliface - Intelligent Assistant for Interfacing Diagnosis and Planning Systems

Status: Completed

Start Date: 2012-04-30

End Date: 2014-08-29

Description: To integrate automated diagnosis and automated planning functions, one must translate diagnosed system faults to corresponding changes in resource availabilities. Implementing reliable translation is challenging, time-consuming, and error prone. We propose to develop Intelliface, an intelligent tool for developing interfaces between diagnosis and planning systems. Intelliface will help ensure that plans are revised appropriately when faults occur in complex space systems. In addition, Intelliface will reduce the effort needed to integrate diagnosis and planning systems. Intelliface will encode and apply a qualitative understanding of generic types of devices and their underlying physics (e.g., electrical storage, distribution, and consumption; fluid flow and storage; signal processing, etc.) in order to identify each activity's direct and indirect resource requirements and their dependencies. Intelliface will use the results of this reasoning to generate resource declarations, updated resource availabilities, and some planning constraints in the planning domain modeling language. In addition, Intelliface will support NASA's top-down systems engineering processes for specifying system functional requirements, performance requirements, and interfaces at each system tier. During Phase 2, we will develop a technology readiness level 6 software prototype that demonstrates the feasibility, utility, and usability of the Intelliface concept within a NASA-relevant environment.
Benefits: Promising non-NASA applications include the management of commercial and government computer systems and communication networks that are used to process large, scheduled computing and data transfer tasks. Management of scientific computing clusters might benefit from Intelliface technology because the servers at various scientific computing centers have different capabilities due to differing hardware, software, and data. Also, batch scientific computing jobs are resource-intensive, so they are scheduled individually.

The primary target applications will be the integration of diagnosis and planning functions for NASA manned and unmanned missions, technology studies such as Desert RATS, and resource management of NASA satellite network resources. Improved integration of diagnosis and planning functions is most valuable when the translation from diagnostic states to reduced planning resources and their impact on plans is complex.

Lead Organization: Stottler Henke Associates, Inc.