Advanced Minehunting Sonar Data Fusion

Arete Associates proposes to develop a fusion algorithm (SONAR FILM) utilizing the existing volume and side-scan SONARs available on the Navys Minehunting Sonars (AN/AQS-20C and AN/AQS-24) to improve detection localization and reduce false detection rates. Our approach builds upon over 30 years of exploiting powerful data analytic tools and previous work with the AN/AQS-20. Using Arete's in-house developed tools, each detection will be localized in 3D space utilizing measurements made by the different modalities at different times. A modality specific feature vector will be used to classify mines, tethers, and anchors for moored mines tethered to the sea floor. The results of the localization and classification will be used to identify different detections as being part of the same mine and fuse them into one feature vector. Finally, Arete-developed framework will be used to classify each fused feature set as being either target or false alarm. The optimal combination of spatial and temporal information from each sensor will provide a powerful and robust object classification algorithm that will reduce false alarm rates and improve localization in SONAR MCM applications.