 |
 |
 |
 |
| |
|||||
|
|
|
Home > News & Events > MITRE Publications > The MITRE Digest > | |||||||||||||||||||
Building Dynamic Web-Based Tools for Tomorrow's Battlefield June 2007
A convoy of U.S. Army vehicles motors through the desert, ferrying supplies to troops battling insurgents. The convoy's long-range air-to-ground surveillance system detects a ground moving target, but doesn't detail its type or origin. The soldiers behind the wheel face a difficult choice: stay on track or seek an alternate route? Groundbreaking research aimed at beefing up traditional command and control (C2) systems with next-generation "Semantic Web" capabilities, now underway at MITRE, aims to dramatically improve warfighters' decision-making capabilities in such situations. MITRE's work in developing ontologies, semantic rules, and "automated reasoners" (models and tools that help computers parse and interpret the meaning of words and images) for Web-based C2 systems will allow the systems to automatically modify the information they process in response to changing environments and events. This MITRE-sponsored research project is forward-looking in more ways than one. It anticipates a future in which dynamic events such as sudden sandstorms or the discovery of unidentified obstacles can trigger the application of more conservative rules of engagement, as in this example, ensuring a continued safe course for the convoy. But beyond that, it envisions a world in which the so-called Semantic Web—now still largely theoretical—is a reality, allowing for more sophisticated machine-to-machine interaction than ever before. "The introduction of ontologies and semantic rules allows us to represent the meaning of terms in a formal way, so that machines can read, compose, and interpret those meanings in a way that's comparable to a human," explains Leo J. Obrst, a MITRE artificial intelligence engineer, theoretical and computational linguistics expert, and group leader of MITRE's 12-member Information Semantics team. By applying emerging Semantic Web technology to today's C2 systems, intelligence and other types of battlefield data can be parsed, interpreted, and transferred to soldiers in real time. Critical decisions such as last-minute course changes or corrections will be supported by vastly improved situational awareness—a feat made possible by systems that can "reason," at least on a rudimentary level. "The vision is that people and computers work together using Web resources," adds Suzette K. Stoutenburg, a MITRE information systems engineer and co-lead on the research project team. Semantic Web Will Transform Military Systems The eventual impact on military systems is hard to overstate, given the sea change that the Semantic Web's adoption will represent, according to Obrst. "We're moving from a keyword search model to a semantic search model. With Google, you search for words, or strings—sequences of characters. With semantic technologies, you search for meaning," he explains. The Semantic Web, ultimately envisioned as an evolution of the World Wide Web, will enable searches that can tell users how close the results are to what they need, how different the results are, and how users can potentially use the "close" results for their intended purpose.
He illustrates the importance of semantics to information technology in the following way: "Humans have very complex knowledge of how things work in the world—for example, you can't put an elephant on a table—but machines don't understand this." By implementing semantic rules in C2 systems, such advanced understanding by machines moves closer to becoming a reality. While the existing Web is designed to present information to people, the Semantic Web will make information meaningful to machines by attaching well-defined meanings, or semantics, to the information. Semantics are expressed in what are called ontologies—frameworks of definitions of terms and concepts that flesh out the relationships of those terms and concepts to one another. MITRE's Semantic Web rules project team has developed ontologies focused on convoy movement. "In theater, you get hostile objects such as snipers, improvised explosive devices, or suicide bombers," Obrst says. The ontologies provide rules for determining what a convoy commander should do under certain conditions. Building Machines that Can Reason Through the ontologies, varying options are built into C2 systems—"It could be 'stop,' 'speed up,' 'change direction,' or 'go into defensive posture,'" Obrst explains. "The ontologies represent behavior and policy rules so that the machine is reasoning about the alternatives. New information comes in, and the machine registers that and reacts accordingly. The commander still ultimately makes the decision, but the machine provides input. It's interacting at a human level." In such interactions, unmanned aerial vehicles could be sent to do reconnaissance and surveillance, their paths dynamically adjusted as conditions change. Requests for backup and troop reallocation could be automatically generated in emergencies. Requests to clarify information between components could be exchanged by applications as they monitor battlefield events, resulting in more meaningful alerts for warfighters. Another potential application is in processing real-time streaming video for use in improving situational awareness. By comparing successive images using complex rules, visual information can be integrated into a comprehensive knowledge base. The three-year research project, now in its final year, has resulted in the identification of critical sponsor requirements for semantic standards and the production of a "reasoning engine" that operates using flexible ontologies, rules, and open standards. The project is now focused on addressing forward-looking issues such as ontology lifecycle management and contextual reasoning. Resolving Ambiguity The flexibility of rules is important, Obrst says, because often information must be exchanged between numerous organizations, such as coalition partners. Many of these different groups don't use the same terminology for similar things. For instance, altitude may be measured from the center or surface of the earth, in meters or miles. By applying rules as information is exchanged, data exchange is enhanced. Because coalition partners may contribute different capabilities to a theater, facts could be exchanged and reasoned over, resulting in improved situational awareness for allied organizations. And information sharing rules can be based on evolving events, so that updates are sent to coalition partners on a dynamic basis. In addition to this Semantic Web rules research, the Information Semantics team also is involved in additional work using semantic technologies. This work includes the development of "taxonomies," or classifications, for various terms, or vocabularies, in military and government domains, says MITRE researcher Ken Laskey, a lead artificial intelligence engineer. "We work with data to get real-world effects, because different data speak different languages," he notes, citing as an example words with ambiguous meanings, such as "target." "Is the target something to be destroyed or protected? What we're trying to do is capture meaning so that when a system user finds a resource he thinks will help him, he truly understands what that resource is." This is not only useful but necessary for emerging technologies such as Web services and Service-Oriented Architectures. Defining Terms for Military Sponsors Building on this and related work, Obrst's Information Semantics team is now also working with the Air Force to develop a "core taxonomy," or set of defined categories, and community of interest vocabularies and models for extending semantic search capabilities to various Air Force technologies. The Department of Defense (DoD) already has a core taxonomy, Obrst says, and the individual military services are working on their own taxonomies and vocabularies to link to the DoD's. "The idea is that communities of interest come up with their own vocabularies and models, and they are linked to the core taxonomy," or master list of categories, he explains. The result will be smoother data interchange between the military services, other government agencies, and coalition partners. Such projects represent the next wave in extending net-centricity to the battlefield—and, by extension, adding a layer of security for soldiers, Stoutenburg notes. "Ultimately, this is about producing agile systems," she says. "The systems will perform many of the functions done today by humans." —by Maria S. Lee Related Information Articles and News Technical Papers and Presentations
Websites |
||||||||||||||||||||
| Page last updated: June 14, 2007 | Top of page |
Solutions That Make a Difference.® |
|
|
