SBIR-STTR Award

The Battle Force Tactical Training (BFTT) system provides realistic, high-stress combat team training. Current implementation relies upon pre-scripted simulation in the synthetic battlespace, resulting in rigid and non-interactive behavior. JSAF models have been integrated into BFTT to begin addressing limitations. However, BFTT must be further enhanced to 1) provide an operator interface to facilitate dynamic instantiation, monitoring and tasking ISF entities; and 2) improve representation and reasoning capabilities of ISF entities. This paper proposes developing a "distributed reasoning" architecture to support new behavioral component integration and take advantage of technologies and approaches not supported by JSAF. These "plug-able" behavioral entities will be attached to BFTT via an HLA "bus" and will drive existing JSAF physical models via the SAF Generic Model Interface. Exploring this new architecture and as proof of concept, a well-understood/simple behavioral model will be re-implemented using LISA, an open source platform supporting intelligent software agent development. Finally, to support operator interaction, a dynamic, enhanced representation of entity subsystems will be developed. This capability will allow BFTT to 1) observe or manipulate a synthetic entity's reasoning and behavioral models interactively; and 2) add/remove behaviors during runtime. The proposed architecture will significantly enhance the combat training environment. A distributed reasoning environment that leverages existing JSAF physical models provides an opportunity to develop more powerful and sophisticated behavioral and reasoning components that can take advantage of a wide array of applied AI technologies.

The US Navy?fs Battle Force Tactical Training (BFTT) system is currently being deployed aboard US Navy surface combatant ships to provide coordinated, realistic, proficiency training. In the current implementation, the fidelity of the model representations is not sufficient to achieve all training objectives. Additionally, the constituents of the synthetic battlespace and their behaviors are based upon pre-scripted simulation scenarios and the system has limited capabilities to interactively and dynamically modify the run-time state of the simulated environment. The Joint Semi-Automated Forces (JSAF) Federation has been successfully integrated and used within BFTT and has been identified as one of the key applications to improve upon the existing BFTT simulation technology. This Phase II SBIR will enhance the capabilities demonstrated in Phase I in the following manner: ?í Extend JSAF?fs distributed protocol to support a broader set of configuration and control directives ?í Integrate the Phase II capabilities into the BFTT Scenario Generation and Control (SG&C) architecture ?í Conduct a technology demonstration(s) to elicit feedback from fleet representatives regarding Phase II capabilities The activities described above will significantly reduce the workload of the BFTT operator while enhancing the training that the warfighter receives. The voice component of the proposed Phase II system will provide a practical application of cutting-edge voice technologies. Lessons learned from Phase II activities will be provided to any applicable vendors in anticipation that technology improvements will be made. The grammar that is used in conjunction with the voice recognition and generation systems will provide a real-world testbed for these technologies. The protocol and the interface module that are being developed for Phase II will be capable of supporting differing simulation domains and intended uses. Experimentation, analysis, and acquisition communities can reuse the simulation technologies that are embedded within these applications in order to understand their individual program objectives.

Keywords:
Intelligent Synthetic Forces, Jsaf, Modeling And Simulation, Bftt, Training