Cognitive Decision Aid Knowledge Acquisition Toolkit (CDAKAT)

"The Rotorcraft Pilot's Associate demonstrated the utility of Cognitive Decision Aiding Systems (CDASs). Unfortunately, the knowledge required to build CDASs is a tremendous hurdle to their de-velopment, fielding and maintenance. In Phase I, we documentedthe largest hurdle to efficient CDAS knowledge management as the `gap' between Knowledge Acquisition (KA) and Software De-velopment (SD). Technology exists to capture knowledge; putting it into a useful, accessible format for later SD lags behind. InPhase II, we will implement and test our prototype solution: TAGGER. TAGGER will enable richer communication between KA and SD by allowing `tagging' knowledge as it is acquired, using SD-relevant terms. Tagging will be done during knowledge elicitationand will require little extra effort beyond taking notes. Because tags are SD-relevant, later knowledge inspec-tion is facilitated and links are carried forward even into testing and re-design. Autonomous software `critics' will use tagging structures todetect flaws and patterns of interest or concern. We are using the industry standard Unified Modeling Language (UML) to structure `tags' and integrate them into SD, and illustrating UML's ability to duplicate the Task Network Toolset used on RPA andproposed for AATD's Hunter Standoff Killer Team (HSKT). In this effort, we will refine the TAGGER repre-sentation, design user interfaces for the TAGGER tool and prototype a tagging enviro