TDI NOVUS INC — Department of Defense SBIR Phase I: N231-028

Description

Navy vessels rely on proper functioning of their Hull, Mechanical, and Electrical (HM&E) systems to perform their missions. However, faults and errors can occur unexpectedly, causing downtime and potentially hazardous situations. To address this issue, TDI Novus proposes PROPEL to detect, analyze, and classify faults and errors in real-time using probabilistic fault models trained on historical sensor data. The system can provide timely alerts to the autonomy software, off-board operators, or onboard engineers. Further, PROPEL autonomously creates resolution strategies to correct for these faults. The use of machine learning techniques enables the system to continuously improve its performance and accuracy over time. Once a fault has been detected, the system can autonomously find a plan of action to resolve the issue using a self-improving automated planning agent. The agent takes into account various factors such as the severity of the fault, availability of resources, and operational constraints to come up with the best course of action. Additionally, PROPEL can continuously learn new plans using an in-system simulation process, which allows it to adapt to changing conditions and improve its decision-making capabilities. The combination of real-time fault detection, automated planning, and continuous learning makes this system a valuable tool for maintaining the operational readiness of Navy ships.