Adaptable, Automated Troubleshooting Expert

Built-in test, systematic troubleshooting, and tech manuals offer an incremental step towards alleviating troubleshooting complexity, but the ability of system designers to provide meaningful test data and specify optimal troubleshooting at point of design is constrained by necessary development limitations and resources required to execute exhaustive testing and validation of diagnostic strategies. Impact Technologies, in collaboration with the Georgia Institute of Technology, proposes to develop an Adaptable, Automated Troubleshooting Expert decision support tool for the Air Force. The proposed tool leverages existing data sources, from troubleshooting manuals to test and diagnostic data, creating an automated diagnostic expert for advanced troubleshooting assistance and assessment of complex systems. Incorporating maintainer feedback and innovative learning techniques, the tool learns and adapts over time. Much like the grease smudged and dog-eared pages of a good technician’s manual, the learning algorithm captures knowledge and preserves and extends its utility for broader use. Much like a “learned expert” or “esteemed colleague,” the intelligent, expert diagnostic support system provides a time and cost saving resource for maintenance and support of Air Force systems ranging from mechanical to electrical and electronic. The proposed modular software architecture and standards-based, open, interfaces support adaptability and reusability across multiple application domains. BENEFIT: The intelligent diagnostic decision support tool offers reduced overhead associated with long troubleshooting times for failures in complex systems, along with a potential reduction in required technical expertise and manpower. Automated learning and encapsulation of institutional knowledge extends the utility of local expertise. Commercial applications include transportation, medical, manufacturing, and utility domains. Complex systems must be maintained and the technology boom has created increasingly complex systems and a need for this technology.