The Automated Logistics Environment (ALE) records all bus communications, systems sensors, and built-in test (BIT) data on every E-2D flight. The ALE today does not leverage the full power of the data available. There are several areas where employing the most up-to-date machine learning and data visualization techniques to analyze ALE data could be used to vastly improve readiness. These techniques allow for easy comparisons of performance metrics for an individual aircraft in the context of its recent history; in the context of peer aircraft with similar usage and maintenance history; and in the context of the entire fleet of aircraft. Additionally, the BIT data can be transformed to be both more readily understood by humans (e.g., visualizing trends) and machine learning techniques. Finally, the fusion of all sources of data (including text, BIT, and sensor) related to the aircraft allows for a comprehensive view of asset health.
Benefit: These new capabilities both have the potential of increasing readiness and make a policy of condition-based maintenance based on ALE data a realistic one. This way maintenance is performed only as needed. Another benefit to predicting failures is to stop disabling failures and fatal failures before they occur.
Keywords: Deep Learning, Deep Learning, maintenance, Logistics, ALE., Machine Learning, Aircraft
