1 American Institute of Aeronautics and Astronautics

Transcription

1 Lackey, S.J., Merket, D.M., Stacy, W., and Freeman, J. (2004). Intelligent training support tools: technology for the future. Proceedings of the 2004 Conference of the (AIAA). 1

2 Intelligent Training Support Tools: Technology for the Future Stephanie J. Lackey, Computer Engineer, and Danielle M. Merket, Research Psychologist Naval Air Systems Command (NAVAIR), Training Systems Division, Orlando, FL, Webb Stacy, Vice President, Technology Aptima, Inc., Woburn, MA, and Jared Freeman, Vice President, Research Aptima, Inc., Washington, DC, Naval aviation has historically led the military training community in the field of modeling and simulation. Most research and development in this area has focused on hardware and software solutions to address issues such as visual fidelity and physics-based modeling. However, there is a clear need to integrate principles of learning with hardware and software solutions for tools to enhance training devices. Modeling and simulation techniques, specifically Object-Oriented (OO) Intelligent s (IA), provide technological advances well suited for assisting instructors in accomplishing training goals. I. Introduction HE Naval aviation community faces evolving training challenges. Legacy simulation systems were not T developed to support a Distributed Mission Training (DMT) environment, nor are they suited to support growing DMT requirements. In particular, legacy systems typically address the operations of a single platform, and they optimize physical fidelity rather than instructional effectiveness 1. The Navy Aviation Simulation Master Plan (NASMP) lists requirements that address these issues. The Air Warfare Training Development (AWTD) program offers promising solutions to NASMP s requirements. AWTD is an advanced R&D program underway in the U.S. Navy that investigates, demonstrates, and integrates strategies and technologies for distributed and deployable simulation-based aviation training. This effort, sponsored through the Naval Air Systems Command (NAVAIR) Program Manager Aviation-205 (PMA-205), focuses on rapidly transitioning mature technologies to the Fleet by developing prototype products for specific simulation and training applications. In addition, this effort investigates elements that have been called out by acquisition efforts (e.g., NASMP) as areas in need of R&D for future transitions as well as strategies and technologies that can be transitioned into existing platforms in the short-term. Two areas of investigation within AWTD include the Common Distributed Mission Training Station (C-DMTS) and Intelligent Training Support Tools (ITST). The first thread, C-DMTS, aims to improve simulator control stations used by instructor/operators by developing a common framework for a multitude of platforms 1. The second thread, ITST, supports the development of a C-DMTS by pursuing tools and strategies related to distributed performance measurement and debrief preparation. The remainder of this paper will describe the C-DMTS and ITST work, and discuss how OO techniques facilitate this effort. II. C-DMTS Background The role of this type of military training research is to bridge the gap between emerging, yet proven strategies, techniques, and technology in the operational environment. AWTD and its subsidiary efforts, C-DMTS and ITST, aim to improve the aviation training environment by providing tools for instructors that capitalize upon cutting edge technology blended with proven training methodologies. 2

3 Due to the movement toward distributed training, and the issues related to legacy systems, it is apparent that new simulators are required. The C-DMTS effort proposes a common framework (Fig. 1) to be used across a multitude of platforms based on common features identified. Each platform could then extend the common framework to meet platform specific needs. The benefits to such an approach include: 1) reduced cognitive workload for instructors, 2) improved subsystem development, enhancement, and interoperability, 3) reduced system acquisition costs, 3) reduced training time and resources necessary for initial DMTS training, and 4) increased use of simulators due to improved feature utilization 1. Syllabus T&R Matrices Wing Training Manuals NATOPs Learning Management System Track Student Electronic Training Jacket Mishap Database Previous Reports of Problems Scenario Generation Events Malfunctions Crash/Ground Control CGFs Exercise Control Monitoring Displays Environmental Malfunctions Realtime Malfunction Insertion Instructor Flags Freeze/Unfreeze Crash/Grounding Control Instructor Operator Station IOS Instructional Sub-Systems Student Planning Instructor Aids Help Files PDAs Brief Presentation Capability Maps Conferencing Debrief God s Eye View Instrument Data Measurement Measurement Reports Automated MOEs/MOPs Event Marking Data from Other Platforms Technical Operating Station (TOS) Displays/Controls to check training device subsystems health, reset programs, & initialization Figure 1: The C-DMTS conceptual model. Scenario development, exercise control, performance measurement, and brief/debrief systems comprise the primary instructional systems required by the C-DMTS. ITST contributes to scenario development and brief/debrief capabilities, but most directly impacts performance measurement. C-DMTS guidelines 1,2 define performance measurement as, Provides for the establishment, collection, and processing of participant performance data into a format usable by and meaningful to the instructor in evaluating performance. This feature reduces the multitude of data generated to only those variables relevant for training the task(s) and works in conjunction with the scenario development and exercise control subsystems. This system also aids in the debrief preparation (e.g., providing performance reports to the instructor.) This subsystem includes all tools the instructor has for monitoring and measuring performance. This includes both automatic and manual measurement and analysis capabilities. 3

4 III. Facilitating Measurement in C-DMTS through ITST ITST focuses on delivering a suite of intelligent agents (IA) currently available to assist instructors by 1) capturing and storing data that is too difficult or not possible for a human instructor to extract from a simulated flight scenario, 2) configuring data for performance measurement and computation, and 3) providing rudimentary debriefing and agent visualization materials 3. Figure 2 depicts ITST s IA architecture. Assessment Diagnosis Coaching Debriefing Debriefing Interface Data Store Measurement s Data Configuration Measurement Computation Mission Measurement Model Configuration Interface Other Data Sources Survey Domain Expert Data Capture HLA Bus Controller Console Student Controller Figure 2: ITST agent architecture. Although many of the agent names clearly indicate their purpose or role, some deserve greater attention. The Data Capture listens for pertinent data available on the HLA network during the scenario execution and forwards it to the Data Store. This storage capability maintains the relevant data from the Domain Expert, and when appropriate, the Other Data Sources. The Domain Expert provides the expected performance data used in the comparison of actual trainee performance to expected performance. The Other Data Sources contains raw data and other data used in the computation of performance measures. The core of this suite of agents resides in the Measurement, which consists of three components: Data Configuration, Measurement Computation, and Mission Measurement. The Measurement agent provides two services. First, the Data Configuration configures the mission performance measures (these parameters comprise the Mission Measurement Model) to support the Premission phase. Second, after an exercise, the Measurement Computation and the information contained in the Mission Measurement Model are used to compute the desired performance measures. The output supports the After Action Review phase of training, by providing valuable data in a format useful to instructors. The result is a greater amount of diagnostic performance measurement and analysis at the fingertips of the instructors that was previously unavailable. IV. Measurement Objects and the Benefits of OO Techniques The ITST effort advances the science of IA technology through the innovation of Measurement Objects (PMO). PMOs, software objects used to measure trainee performance in a federated High Level 4

5 Architecture (HLA) simulation environment, expand the previous world of objects in an HLA federation. In particular, PMOs provide a meaningful yet precise description of measurments in the simulation system, affording the human operator/instructor the opportunity to interact with it in intuitive and useful ways 3. AWTD-ITST enhances the state-of-the-art in IA technology and represents an instantiation of future technology. s within this system are implemented as objects, as are communications between them. The agents and communication objects function as other HLA objects, which emulate aspects of the physical environment such as aircrafts, ships, bombs, or radars. In fact, the communication objects expand the typical world of HLA objects. 3 Advantages of this approach include reusability and extensibility. Reusability is not only a theoretical advantage, but also a focus of the Chief of Naval Operations (CNO) 4. The realities of Navy training conditions and requirements calls for blended solutions to achieve individualized learning based on reusable learning objects. Virtual reality and intelligent agents capitalizing on object-oriented techniques are a natural fit for the type of solution sought by the CNO. The agents built under the ITST program have been specifically designed and implemented as extensible objects to facilitate future work. With slight modification, the agents built for the aviation domain in ITST could be applied to a number of military domains, such as combat information center operations. Though seemingly unrelated, aviation and CIC operations both require distributed teamwork, and are often part of the same air warfare organization. Another powerful benefit of ITST s OO approach is the encapsulation of private instance variables that describe the instance states and instance specific information needed by the respective objects. By hiding complexity behind the individual interfaces, future implementation changes will result in minimal impact on the code using the respective object(s). Encapsulation provides for improved quality of code and speed of development. Finally, ITST s approach represents training-relevant information at the level of trainees, teams, and measures rather than column-order arrays, hash tables, and string manipulation. The result is clearer thinking and faster solutions to new and emerging requirements. While object-orientated techniques do not literally represent categories the same way as humans do, they do give developers the opportunity to think at a higher level and benefits the speed and quality of implementation. V. Future Efforts Integration of this suite of agents into a distributed multi-platform testbed representing the E-2C Hawkeye and F/A-18 Hornet is scheduled for Fall of Future agent work is planned to address performance assessment and diagnosis, coaching, advanced debriefing, and scenario generation support. VI. Conclusion The classic advantages of OO techniques 5 are clearly part of the IA architecture implemented by ITST within the C-DMTS framework. OO techniques facilitate the integration of ITST agents into the C-DMTS framework. The ITST agents benefit from the extensibility of OO, and offer their functionality to other domains through the reusability inherent to their design. Applying these advanced technologies to simulated aviation training afford new opportunities to improve warfighter performance in the operational environment. Acknowledgments The authors would like to acknowledge NAVAIR PMA-205 and the Manned Flight Simulation Division for continued guidance and support in this endeavor. The authors would also like to recognize the collaboration efforts of the AWTD C-DMTS team from NAVAIR Training Systems Division. References 1 Walwanis Nelson, M.M, Smith, D.G., Owens, J.M., and Bergondy-Wilhelm, M.L., A Common Instructor Operator Station Framework: Enhanced Usability and Instruction Capabilities, Proceedings of the Interservice/Industry Training, Simulation, and Education Conference, Orlando, FL, Walwanis Nelson, M.M., Smith, D.G., Owens, J.M., Stubbs, E.C., and Bergondy-Wilhelm, M.L., Enhancing Naval Aviation Simulator Instructional Tools Through Psychological Research, Poster presented at the 49 th annual Meeting of the Southeastern Psychological Association, New Orleans, LA,

6 3 Stacy, W., Freeman, J., Lackey, S.J., and Merket, D.M., Enhancing Simulation-based Training with Measurement Objects, Proceedings of the Interservice/Industry Training, Simulation, and Education Conference, Orlando, FL, in review. 4 Clark, V., Accelerating Our Advantages, Chief of Naval Operations, CNO Guidance 2004, Garrido, J.M. Object-oriented Discrete Event Simulation with Java: A Practical Introduction, Kluwer Academic/Plenum Publishers, New York,