Dynamically Generated, Articulated 3-Dimensional Training Content

The Navy increasingly relies on immersive training solutions utilizing Virtual, Augmented, and Mixed Reality (VR/AR/MR) to provide instructors with the flexibility to create realistic training scenarios, including fault injections and novel situations. However, the traditional development of interactive 3-D models remains a significant bottleneck, requiring months of intensive design, modeling, and coded scripting. This slow process hinders curriculum updates, constrains training throughput, and increases lifecycle costs. Navy initiatives demand technologies that accelerate content creation and refresh, particularly given emerging opportunities for advanced LVC training, adaptive tutoring, and rapid prototyping of virtual content. Recent technological breakthroughs, such as neural radiance fields, 3-D Gaussian Splatting, and 3-D model diffusion, offer near-instant scene creation and real-time asset generation. Furthermore, advances in computing power and technology such as generative artificial intelligence (AI) enables the ingestion of large technical datasets to auto-draft procedures and annotations for instructional content. Despite these advances, existing solutions often lack the dynamic articulation, physics, and logic necessary for hands-on procedural practice and team interaction. Critically, fleet use cases require no-code tools that empower Subject Matter Experts (SMEs) to verify content outputs, embed adaptive cues, and seamlessly export to existing simulators and LVC systems. Dynamic articulation, multi-user networking, SME oversight, and standards-compliant export represent key transition hurdles. Secure, user-friendly toolkit solutions are desired that enables the rapid conversion of diverse Navy data sources (structured and unstructured) into photorealistic 3-D assets suitable for deployment in AR and MR training platforms and LVC environments. The solution should provide a low-code/no-code software platform for the automated generation of AR/MR training content that requires minimal source media (e.g., smartphone photos, video, Computer-aided design model, or schematics) while offering human-on-the-loop controls for validation and refinement of models. These photorealistic 3-D assets should have articulating sub-components, respond to user actions, and drive the learning of procedural maintenance workflows. The toolchain must support easy export to a variety of Navy immersive reality training platforms and/or LVC environments or systems. The goal is to produce mission-specific, realistic content in days instead of weeks or months, leveraging actual fleet data and user-friendly tools. This topic seeks an innovative capability that supports automated or semi-automated generation of AR/MR content from real-world data sources—such as maintenance records, after-action reports, sensor logs, or mission debriefs—and enables curriculum developers or instructors to generate varied training scenarios quickly and intuitively. The system should support parameterized scenario generation and include tools for editing or customizing scenarios without requiring software development skills. Key technical elements may include: • AI-assisted generation of 3-D assets, animations, or procedural scripts; • Natural language processing (NLP) to parse textual data into scenario logic; • Interfaces that allow low-code or no-code authoring of AR/MR content; • Export capability to common AR/MR engines (Unity, Unreal) and Navy-deployed HMDs or mobile devices; and • Security and compatibility with Navy computing environments. The expected outcome is a toolchain that empowers training programs to rapidly create and deploy operationally relevant AR/MR training content, significantly reducing development time and cost while enhancing the effectiveness of Navy training programs.