Autonomous Multi-Intelligence Acoustic Reasoning Agent for Anti-Submarine Warfare

SBIR topic DON27BZ02-NV014 (U.S. Department of Defense – Navy), Phase I and II. Open to U.S. small businesses. Opens: December 2, 2026. Closes: January 6, 2027. The Department of the Navy (DON) seeks development of an autonomous decision-support and reasoning capability for ASW that can reduce dependence on continuous human interpretation of complex acoustic displays and enable future distributed unmanned ASW operations. Current ASW performance is heavily dependent on trained operators who interpret passive narrowband, passive broadband, Demodulation of Noise (DEMON), and other acoustic products while simultaneously incorporating situational knowledge such as geography, bathymetry, Conductivity Temperature and Depth (CTD) data for sound-speed structure, Global Positioning System (GPS), Automatic Identification System (AIS), organic/non-organic tracks, surface traffic, tactical geometry, and expected target behaviors. This human fusion process is effective but difficult to scale across persistent unmanned systems operating at range, under intermittent communications, and in acoustically cluttered environments. This SBIR topic requests development of an autonomous ASW agent that can ingest heterogeneous acoustic detections and associated confidence measures, correlate them with non-acoustic common operating picture data, reason over environmental acoustic propagation, and maintain multiple explainable track and classification hypotheses over time. The capability should explicitly account for confusers such as merchant shipping, fishing vessels, biologics, own-force and friendly surface ship contributions, reverberation/clutter, and geographically-dependent propagation effects. The resulting system should not be a single-sensor classifier only; it should demonstrate multi-source temporal reasoning that can explain why a contact is assessed as likely submarine, non-submarine, ambiguous, or requiring additional sensing. The approach should be modular and compatible…