Current needs in tactical situational awareness require a new type of infrastructure to encode, transmit, store, fuse, and
display vastly heterogeneous data that may include "hard" sensor types including video, radar, multispectral, acoustic
sensor array, 3D flash LIDAR, and "soft" sensor inputs such as textual reports from trained and untrained personnel,
unsolicited and solicited open source web information, and hybrid "hard/soft" data such as human-annotated image or
video data - which can be highly useful, but difficult to categorize and exploit.
While the demand for scalability, rapid deployment, and decentralized access to data and services grows, the need for
data security and integrity is as critical as ever. Methods for handling the conflicting needs between access and security
are addressed. Furthermore, the evolving role of humans in data fusion systems must be addressed by the infrastructure.
In addition to systems enhancing human data analysis capabilities through advanced visualization and sonification
techniques, the data itself is more likely to contain information about humans - which is not always a task well suited to
conventional data storage and retrieval methods.
This paper describes a multi-agent approach to designing a secure, distributed, service-oriented infrastructure to support
human-centric hard and soft information fusion.