SBIR-STTR Award

Synthetic scenario-based training of Navy personnel in the use of Navy SIGINT/IO systems has helped to reduce training costs, and it has enabled the personnel to be trained in an environment that sufficiently approximates real-world situations that could not otherwise be accomplished within the class room. However, scenario development is highly complex and involves a great deal of human effort and domain knowledge, discouraging the modification of existing scenarios to keep them current in an ever-changing threat environment. This problem is exacerbated when the scenario represents a combination of multiple data sources. The proposed research will show that the use of static models and companion correlation modules during scenario creation will reduce the complexity of scenario development and reduce the domain knowledge required. Static models can be devised to encapsulate domain knowledge for a particular data source, and correlation can be used to fuse the output from each static model to produce a cohesive scenario. To enable autonomic generation and regeneration of multi-source scenarios, the proposed research will also address service composibility and data heterogeneity among the participating static models and correlation modules. Finally, the proposed research will investigate human-computer interfaces for guiding the scenario developer through the autonomic process.

Benefit:
Scenario based training through virtual simulation and virtual stimulation has wide marketability. Any sophisticated computer-based system can benefit from scenario-based training. Scenario-based training is especially important for systems that are involved in life and death situations such as first responder incident management systems, critical infrastructure systems such as harbor security and traffic management systems, and national security systems such as DoD systems and border patrol systems.

Keywords:
Distributed Mission Training , Distributed Mission Training , Virtual & Constructive Training, team training,, Tailored Training, LIVE, training exercises

Synthetic scenario-based training of Navy personnel in the use of Navy SIGINT/IO systems has helped to reduce training costs, and it has enabled the personnel to be trained in an environment that sufficiently approximates real-world situations that could not otherwise be accomplished within the classroom. However, scenario development is highly complex and involves a great deal of human effort and domain knowledge, discouraging the modification of existing scenarios to keep them current in an ever-changing threat environment. This problem is exacerbated when the scenario represents a combination of multiple data sources. The proposed Phase II effort will leverage the positive results from the Phase I research to develop a fieldable Scenario Generator able to output Stallion-ready training scenarios. The Scenario Generator will make use of data-driven static models developed during Phase I, which significantly reduced scenario creation time and reduced the domain knowledge required. The Phase I research showed that domain knowledge, encapsulated within selected data source, could be used to drive static models, and that those static models could be orchestrated such that their output produces a cohesive, multiple-Intelligence (Multi-INT) scenario.

Benefit:
The knowledge gained during the Phase I prototyping and experimentation will be leveraged in Phase II. The result of Phase II will be leveraged across a variety of Navy/DoD systems to satisfy training needs. If our approach is successful, we will add fidelity and complexity to training scenarios while at the same time making them more affordable. The end result will be that servicemen and women will be better prepared to do their jobs. Our research has also shown some promise in other areas. The Markov modeling, for example, lends itself to several DoD related needs. Specifically, the developed models could be used to identify state changes based on various scenarios and applied to real-world data. The results could be developed to provide predictive analysis in a variety of circumstances. Additionally, these models have shown some promise in identifying changes in wireless network traffic communications packets. Applying what has been learned during Phase I and what we expect to see in Phase II could help to provide both network/cyber situational awareness and indications and warning of cyber attacks.

Keywords:
synthetic, stallion, scenario, intelligence, Simulation, Training, Navy, Modeling