Situational Awareness for Mission Critical Ship Systems using Probabilistic Knowledge Graph

This effort proposes to develop situational awareness methodologies for mission critical ship system based on the state-of-the-art probabilistic knowledge graph (KG) and deep learning. The proposed KG approach can incorporate various data fusion technologies for analysis of unstructured data (text, images, etc.) and structured data (signal feeds, database items, etc.) for automated decision support and predictive capabilities. Specifically, the effort proposes to design and develop a general and configurable KG framework that can be integrated and applied to specific operational machinery control and condition monitoring systems (MCS). With deep neural network learning integrated in several key components of the system, such as knowledge fusion, pattern discovery, and prioritized action recommendation, the proposed KG-based cognitive framework permits a rich representation and reasoning of the semantics context in the data. The resulting KG MCS products developed under this effort are capable to enhance state and situational awareness of shipboard machinery control operations for real-time decision support. It is expected that the final products could be incorporated and integrated into the algorithmic base of the standardized MCS baseline.