ADAPTIVE METHODS, INC. — Department of Defense SBIR Phase I: N162-093

Description

The Navy is facing more sophisticated Undersea Warfare threats as the technology and tactics of our adversaries improve, and the use of complex multistatic sonobuoy fields offers significant advantages in detection of these threats. Due to the size and complexity of the processing of multistatic fields, more information is generated than sonar Operators can manually evaluate. Adaptive Methods, teamed with RDA Inc., proposes to develop an innovative method of reducing clutter and detection prioritization for software TDAs that organize and highlight high contact probability data. Our approach is multi-fold, first we propose reducing the overall false alarm rate by adding adaptive signal processing to reduce sidelobe leakage caused by acoustic interference. Additionally, we will incorporate Adaptive Pulse Compression classification feature extraction to delineate contacts. The third step is using these extracted features as inputs to a machine-learning algorithm to produce a single contact probability score. The final addition is to adjust these probability scores based upon a kinematic evaluation of the contacts against previously processed data from the sonobuoy field data. We believe the addition of these techniques will reduce the false alarm rate and the TDAs will guide the Operator to the highest quality contact data for review.