Mobile-Agent-Based Autonomous Data Fusion for Distributed Sensors

Autonomous Data Fusion is an essential technology to support the non-traditional surveillance systems that employ small autonomous systems communicating with each other – forming a distributed sensor network. Based on the Phase I achievements, Impact Technologies in collaboration with Wright State University, Virginia Tech and Northrop Grumman, proposes to further develop, mature and demonstrate with field trials a mobile-agent-based decentralized data fusion system that fuses the data from sensors deployed on unmanned vehicles into an integrated ISR information system on the combat ships. The proposed data fusion architecture is a three-level system which performs data processing and fusion at Node, Group, and System levels, respectively. The system has a number of advantages over conventional solutions: 1) communication bandwidth requirement and energy consumption are reduced by sending light-weight mobile-agents instead of sending data between the nodes; 2) progressive decision accuracy is achieved as the mobile-agents traverse the network performing Bayesian data fusion on the nodes; 3) real-time requirements are fulfilled by the Anytime algorithm design; 4) improved reliability; and 5) high scalability. Thorough and extensive experimental tests and field trials will be carried out on heterogeneous USVs to validate the developed technologies and facilitate the technology transition.