SBIR/STTR Award attributes
Research Proposed: Research is proposed to research and design a novel Unmanned Vessel Health Monitoring System (UVHMS), providing an onboard diagnostics and prognostics system for USV and UUV, supporting the identification of active and imminent platform faults. Using onboard sensors and data processing, multi-sensor cross-correlation, pattern recognition and anomaly detection, the UVHMS will essentially replace the human sensors aboard a manned vessel, allowing the USV/UUV to autonomously make data-driven health decisions, while also adding the capability to periodically provide real-time and predictive vessel health data back to remote stakeholders.Problem Statement: Advancements in autonomous operation, artificial intelligence and machine learning have led to proliferation of unmanned vehicles. While these systems provide significant operational and safety advantages to U.S. future warfighters, they cannot rely on human senses to make decisions about their own health. Plan/Process Outline: Research and analysis will be performed investigating multi-sensor data fusion, using deep learning with principle component analysis, to account for non-linear characteristics when multiple disparate sensors are cross-correlated for detection of abnormal situations or recognized problems. The feasibility of the system will be evaluated and demonstrated using algorithm modeling and simulation, sensor data collection, power draw, data flow, and USV/UUV platform integration constraints.