Abstract: The Infantry Immersive Trainer (IIT) is a mixed-reality training system designed to extend training
capabilities for Marines across a wide range of military operations (ROMO) within a single training
environment. This is accomplished with a cutting-edge mix of real and virtual technologies, with a
configurable hardware and software system infrastructure, allowing training scenarios to be quickly
modified to focus on different training objectives based on an incoming trainee group’s specific needs and
goals.
In this paper, we review our lessons learned in developing scenarios for this mixed-reality environment, In
addition to traditional scenario design challenges, the experience involved new challenges focusing mainly
on maintaining a realistic experience at locations where the physical system and the virtual system
converged. Also, the effort involved using front end analysis to drive scenario design, early in the system
development cycle, providing the opportunity for scenario design to inform the configuration of physical
and virtual capabilities to support increased training value and modularity. Finally, we present a summary of
our results of an initial theoretical training effectiveness evaluation for the whole system, which provides
additional insights to the scenario design and system development process. We conclude with a
recommended approach for future mixed-reality scenario design efforts.
Abstract: As the complexity of simulation-based training exercises increases (e.g., distribution of sites, integration of multiple platforms and services), there is a need to advance training technologies such as Instructor Operator Stations (IOSs). Past work to address IOS deficiencies has focused on identifying common functions that meet the needs of multiple Navy aviation platforms (Walwanis Nelson et al., 2003); however, the increased focus on joint training necessitates a shift in development to support a more diverse set of simulation systems. To address the challenge, the authors have developed a multi-domain, meta-thematic analysis approach. This paper provides an overview of this approach, which draws on work in sociotechnical systems to define a framework for data collection. Specifically, an adapted set of sociotechnical systems factors (Vicente, 1999) provides a meta-thematic framework for collecting domain relevant data across multiple domains. Identifying similarities and differences that exist across each factor
provides insight into common functional requirements as well as system functionalities that require flexibility to meet disparate requirements. Additionally, by providing a method that allows for the analysis of multiple teams, this approach provides a mechanism for identifying weak points in the sociotechnical systems environment where the utilization of adaptive systems would be appropriate.
Abstract: Military airspace is increasingly crowded with
traditional aircraft competing with new loitering munitions and
UAVs. Managing the airspace is therefore more challenging,
requiring closer coordination among all the stakeholders. In this
paper, we describe the motivation and design of a knowledge-
based system that attempts to automate aspects of airspace
management, including the detection and resolution of airspace
conflicts. We then describe a formative evaluation of the system
as compared to human performance of the same task, the
evaluation setup, results, and analysis.
Abstract: We are developing an Interactive Story Architecture for
Training (ISAT) that combines the user-adaptive features of an
intelligent tutoring system with the story management capability of a
scenario director to provide a training experience that is tailored to
individual trainee needs—both dramatic and pedagogical. Another
unique contribution of this effort is the development of an authoring
tool that will facilitate the input of dramatic and pedagogical content
by a non-programmer. The envisioned result of ISAT is a seamless
integration of interactive story and individually-guided instruction.
The current ISAT prototype is tightly coupled with the training
simulation and the corresponding domain knowledge. This is in
contrast to a truly modular architecture design that could
accommodate a variety of training needs, domains, and simulations.
Therefore, we propose a core architecture that would be
supplemented by specialized, modular plug-ins to support unique
training-dependent or simulation-dependent needs. In designing this
modular architecture, we have identified several basic questions
about how to develop the modular architecture so that it
appropriately addresses a variety of training contexts and available
simulation tools. We present these questions, and our initial
considerations of them, in this paper.
Abstract: Cognitive Agent-based Real-time Measurement and Assessment (CARMA) is an automated performance assessment methodology for team training applications. This paper describes CARMA particularly focusing on its application to automated performance assessment of team communications. With CARMA, performance assessment is incorporated into the agent’s reasoning, allowing performance assessment to be intelligently adapted to the developing situational context. Examples of our approach are taken from the Virtual Interactive Pattern Environment and Radiocomms Simulator (VIPERS), an on-going Air Force effort developing a simulation-based training system for guided practice and feedback in radio communications and decision making in the Joint Primary Air Training System (JPATS) overhead pattern.