MOSAIC ATM, INC. — Department of Defense SBIR Phase I: N191-032

Description

Ships operate in an environment with a large number of air and surface tracks, where surveillance data may experience gaps in coverage for many reasons. This proposed effort will develop enhancements to be integrated into both current and future shipboard combat systems to provide more operationally useful and realistic predictions of track behavior and location using machine learning (ML) and artificial intelligence (AI) techniques. Through leveraging historical data repositories of surveillance data, and applying ‘big data’ analysis techniques, the algorithms proposed herein will learn behaviors and predict intentions of surface and air tracks. The ML algorithms will detect patterns in the current data that are similar to historical data, whereas the AI components will perform reasoning and inference based on the current battlespace situation, potential track objectives, and physical constraints.The proposed system will also provide a risk characterization for each track. Situational awareness will be significantly enhanced through AI-based computer automation that detects new or unusual characteristics or threatening behaviors of tracks, and then highlights those tracks as contacts of interest (COI). Similarly, tracks that are detected to behave consistently with known and expected behavior can be de-emphasized to the operator, allowing the operator to focus only on COI’s.