CYANDECA: Cyber Anomaly Detection, Classification, and Analysis for Condition Based Monitoring

Navy is developing the concepts and methods to leverage Machine Learning (ML) techniques for the maintenance decision-making on condition-based maintenance plus (CBM+) platform. Effective health monitoring for condition-based and predictive maintenance requires intelligent sensor selection and placement, and context-aware interpretation of sensor data to detect the many possible fault modes. Moreover, deployment and adoption of sensors can potentially expose the interconnected components in the systems to a wide variety of attack vectors. Thus, it is in critical need to develop ML-based cybersecurity resilience solutions on the CBM+ platform for automated monitoring, detection and identification of suspicious or unusual patterns possibly indicating the presence (or prediction) of a cybersecurity threat, vulnerabilities, or system failures. To address this need, Intelligent Automation, Inc., along with the University of Maryland, propose to develop CYANDECA, a Cyber Anomaly Detection, Classification, and Analysis system that can process the information emerging from the CBM system for cybersecurity protection and resiliency. It can automate the change detection in the information patterns harvested by the CBM, classification of the detected anomalies, and threat investigation and risk assessment. The developed concepts and technologies in CYANDECA will significantly enhance fleet performance and readiness.