Artificial Intelligence, Data Fusion, and Mine Recognition System

During present clay mine search operations, targets are detected acoustically using active sonar and identified by analysis of video-based images. Recognition and classification of mine-like targets in highly cluttered environments are difficult tasks even when using today's most technologically advanced systems. Hybrid systems that make use of acoustic and nonacoustic sensors, artificial intelligence (Al), and advanced multisensor fusion technologies are required to provide improved detection and classification performance. To achieve this goal, a design study will be conducted during Phase I of this SBIR contract. Three objectives will be satisfied during the design study phase: first, acoustic and nonacoustic sensor technologies will be Identified and characterized (i.e., report timeliness coverage, etc.) to determine mission applicability; next, hybrid neural network and AI applications will be defined to provide automated detection, classification, and localization estimates to an operationally proven multihypothesis fusion algorithm; finally, the multihypothesis fusion algorithm will be modified to provide final classification and localization estimates using sensor-level detection data from dissimilar types of sensors. During the transition option phase of this program, AI-based intersensor cueing techniques will be implemented and evaluated.