15B.3 AI-Based Earth and Space Observing Digital Twin Prototype

Thursday, 1 February 2024: 2:15 PM
316 (The Baltimore Convention Center)
Lynn Montgomery, Lockheed Martin Space, Boulder, CO; and P. Davis, G. Hoffmann, B. Ballance, J. Stone, J. Lance, M. Fingerhut, A. Linenfelser, T. Kaye, and M. Wear

Lockheed Martin, in partnership with NVIDIA, is developing a prototype for the National Oceanic and Atmospheric Administration (NOAA) that features an innovative, flexible and visually robust environment for displaying geophysical data from NOAA ground and space-based assets and models. The Artificial Intelligence (AI)-based Earth and Space Observing Digital Twin (EO-DT) integrates mature software from Lockheed Martin and NVIDIA to ingest, process, and display various geophysical and space weather data in an immersive digital environment. The agile, scalable EO-DT framework processes observations from secure ingest through product generation and data fusion to product distribution to end users. Our architecture is composed of three functional components: back-end data processing, real-time interactive simulation, and customizable scientific visualization. OpenRosetta3D™ formats, stores, and applies pre-existing AI/Machine Learning fusion and anomaly detection algorithms to data; NVIDIA’s Omniverse connects multiple applications to a real-time environment; and Agatha visualizes data from multiple sensors within an interactive Earth and space platform. Our flexible architecture is designed to incorporate data from additional and future sensors and readily adapts for interoperability with other initiatives through standard interfaces. The flexible and configurable architecture allows users to view fused geophysical data by geographical area, temporal coverage, and vertical level, providing the foundation for the next-generation enterprise ground system.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner