DoD systems, including manned and unmanned platforms, control systems, and communications systems, require prognostic capabilities to increase reliability and mission success, and reduce sustainment costs, unscheduled downtime and critical failures. Developing prognostic capabilities usually involves building extremely detailed, time-consuming and expensive models that only apply to a single platform. Alternatively, more scalable techniques can develop prognostic capabilities quickly with less expensive models but generally have poorer performance and arent flexible enough to integrate information from more detailed models. Metron will bridge the gap between these two paradigms by creating Insight, a flexible suite of tools for post mission analysis, fault detection, diagnosis and prognosis. Insight will enable exploration of platform data, rapid development of low cost models, definition and integration of high fidelity models for critical systems, and fusion of information from multiple sources to increase the accuracy of fault detection and forecasting. Insight will provide engineers and analysts the ability to generate and apply prognostic capabilities in little time, enabling near-term increases in mission reliability. Ultimately, our 0xA0 solution will increase mission reliability, improve operational availability, reduce support labor requirements, and reduce vehicle maintenance costs. Moreover, these tools?will create a generalizable solution for improving reliability, availability, and safety across the unmanned systems domain, including throughout the government and commercial sectors. 0xA0
Benefit: The proposed solution will increase mission reliability by preventing abnormal behavior from going unnoticed. The addition of prognostic capabilities will further increase mission reliability by predicting failures and abnormalities before they occur. When integrated with maintenance and logistics planning, prognostic capabilities enable improvements in operational availability, reductions in support labor requirements and lower maintenance costs. The capabilities produced by this effort can be rapidly adapted to other platforms to realize the same benefits on other fleet assets.
Keywords: Prognostics, AI, Machine Learning, Reliability, PMA, anomaly detection, Visualization, FDD