Multi-source target localization and state estimation for passive sonar arrays

Sentenai proposes to develop a novel deep learning model to perform target localization and state estimation using data from environmental propagation information, ownship-motion, and sensor data from contact followers, to provide a single set of predictions and uncertainties that can be communicated to an operator using a unified display. To enable this, Sentenai will leverage and extend existing components of its technology previously developed to perform multi-source track fusion on air and surface targets in dynamic environments. Additionally, Sentenai will develop a prototype display that demonstrates the unified target tracking capability and enables an operator to interrogate predictions and link back to the original input information that was processed to make a prediction.