SBIR-STTR Award

Chameleon represents a synthesis of standards, frameworks, tools and processes that achieves adaptive system behavior. It does so by keeping a log of user interactions and utilizing an analytics engine to perform state-trace-analysis and probabilistic computations on that information, with the goal of learning a users preferences. The preferences themselves are in the context of the application domain model, which results in the ability to feed the results of the analytic computations back in near-real-time to the application rendering engine. The result is a UI that is built on-the-fly and adapts to the users usage over time. The application domain model is built in terms of a higher level metamodel that allows for a dynamic data store that can evolve over time without requiring schema changes. The result is a dynamic data model, and a system that requires no downtime when SMEs and data modelers make necessary modifications to the application domain. Agents crawl data sources, analyze and align ontologies, then link to and translate data. The architecture itself is adaptable. The analytics engine and preferences module both utilize plug-in architectures because research in adaptive systems is ongoing and new strategies for solving the problems of adaptive systems are inevitable.

Benefit:
The potential applicability of the Chameleon approach to domain modeling, coupled with adaptive UI behavior, is wide-ranging. For example, consider a system containing airline flight information and reservations. Such a system would be utilized by an individual traveler to book a flight. But a travel agent scheduling multiple flights en masse would require more of a power 0x9D interface. The same domain model would be used to display flight status to those awaiting arrival of a traveler, and would also be used by baggage tracking software. Someone in charge of setting prices for various flights might analyze how full the planes are, and how far they are traveling. Maintenance crews would need to know how many miles particular planes have logged in between inspection and servicing. The list goes on. Each different user base has different usage patterns and requires a different UI with access to a different subset of the domain elements. Usage of an ontology-driven dynamic architecture with an adaptive UI would greatly facilitate the process of evolving the domain model in many industries. In fact any industry with a complex, evolving domain and a diverse user community can benefit from an adaptive system such as the proposed Chameleon. In summary, while the Federal Government across the board needs a capability to address the problem in this SBIR, the Services in particular with its migration to net-centric operations, would benefit greatly from the proposed Chameleon effort to support missions ranging from battlefield operations to disaster relief to humanitarian assistance. If successful, Chameleon would support the Navy Satellite Communications (SATCOM) program and other DoD programs by allowing operators to access an ever-expanding pool of data sources and then customize their views to support various and growing mission areas whether they be command and control, air defense, real-time battlefield logistics, counter-IED, joint fires, blue force tracking, terrain analysis, weather, or network operations, to name a few.

Keywords:
Intelligent Agents, Intelligent Agents, Dynamic object model, Model Driven Architecture, Ontology, Delarative Application Metamodel, adaptive user interface

Chameleon is a synthesis of standards, frameworks, tools and processes. In a nutshell, Chameleon utilizes a combination of Semantic Web/Web 2.0 technologies (OWL, RDF, Ajax, REST) layered upon a platform of more venerable enterprise technologies (RDMBS, Java Virtual Machine) into a Model/Ontology Driven Architecture, with the resulting plug-in based UI having the ability to be built on-the-fly. Our synthesis approach allows us to leverage the extensive research and lessons learned by the community rather than proposing development from scratch while still leveraging leading-edge technologies as part of the solution. This dynamic architectural base will provide a responsive, rapid-iteration environment that will facilitate the exploration and application of cutting edge theories in adaptive user interfaces (Bayesian networks/learning classifiers, artificial intelligence, machine learning, etc.) in order to accomplish the goal of providing an adaptive user interface coupled to a complex data model that satisfies a broad user community with varying needs.

Benefit:
The potential applicability of the Chameleon approach to domain modeling, coupled with adaptive UI behavior, is wide-ranging. For example, consider a system containing airline flight information and reservations. Such a system would be utilized by an individual traveler to book a flight. But a travel agent scheduling multiple flights en masse would require more of a power 0x9D interface. The same domain model would be used to display flight status to those awaiting arrival of a traveler, and would also be used by baggage tracking software. Someone in charge of setting prices for various flights might analyze how full the planes are, and how far they are traveling. Maintenance crews would need to know how many miles particular planes have logged in between inspection and servicing. The list goes on. Each different user base has different usage patterns and requires a different UI with access to a different subset of the domain elements. Usage of an ontology-driven dynamic architecture with an adaptive UI would greatly facilitate the process of evolving the domain model in many industries. In fact any industry with a complex, evolving domain and a diverse user community can benefit from an adaptive system such as the proposed Chameleon. In summary, while the Federal Government across the board needs a capability to address the problem in this SBIR, the Services in particular with its migration to net-centric operations, would benefit greatly from the proposed Chameleon effort to support missions ranging from battlefield operations to disaster relief to humanitarian assistance. If successful, Chameleon would support the Navy Satellite Communications (SATCOM) program and other DoD programs by allowing operators to access an ever-expanding pool of data sources and then customize their views to support various and growing mission areas whether they be command and control, air defense, real-time battlefield logistics, counter-IED, joint fires, blue force tracking, terrain analysis, weather, or network operations, to name a few.

Keywords:
Dynamic object model, Ontology, adaptive user interface, Model Driven Architecture, Declarative Application Metamodel, Intelligent Agents