Project

General

Profile

2021 I/ITSEC: STEEL-R Paper

User documentation
11/09/2021

The Army’s Synthetic Training Environment (STE) modernization program’s Training Management Tools require capabilities that objectively measures and evaluates performance over time. Persistent tracking of individual and team performance data enables these tools to better infer proficiency levels, identify strengths and weaknesses, and adaptively tailor coaching and remediation. The STE Experiential Learning for Readiness (STEEL-R) project addresses this requirement by establishing an interconnected system of systems built on open source software and commonly applied data standards.

The STEEL-R team is developing an extensible data strategy that interoperates across Live, Virtual and Constructive environments. It uses real-time processing to translate data sources into meaningful assessments that align to warfighter competency requirements. To demonstrate this concept, STEEL-R leverages tools and methods from the Army’s Generalized Intelligent Framework for Tutoring (GIFT) and the Advanced Distributed Learning (ADL) Initiative’s Total Learning Architecture (TLA) projects. GIFT is used to capture and interpret raw learner data, then TLA standards and business practices are applied to communicate outcomes to a competency management system for
readiness and talent tracking and to a persistent data lake to support decision analytics pipelines.

In this paper, we describe the functional components of the STEEL-R architecture and illustrate it in the context of a Rifle Squad use case, focusing on data flows and processing from the training point-of-need to the Army enterprise cloud. The STEEL-R architecture serves as a reproducible data strategy for STE that can extend cross-service. It aligns evidence-based metrics derived from operational training exercises with established competency frameworks for every echelon and individual role. These frameworks inform the performance metrics and type of data that must be reported to make meaningful inferences on competency proficiency. We will conclude with a discussion on the future capabilities a training management architecture and set of data strategies of this nature can potentially support.

Downloads

21332.pdf (1.92 MB) Goldberg, Ben, 11/09/2021 04:25 PM [D/L : 16]