Space Domain Awareness (SDA) research and engineering by Deepinder Uppal — Chief Technology Officer, U.S. Department of the Interior, and former CTO / EVP of R&D for space and defense technology programs.
The implementation is private. This repository is the public capability overview, research index, and evidence trail for the DEEPSIGHT line of work.
DEEPSIGHT descends from Deep Sight, a Space Domain Awareness platform Deepinder Uppal built and scaled from a Defense Innovation Unit (DIU) accelerator prototype into a deployed production system supporting U.S. Air Force and U.S. Space Force space-situational-awareness mission sets. DIU profiled this commercial SDA work in Commercial Space Industry Providing Tools to Advance and Augment Space Superiority (2022).
The current research platform continues that line: a full-catalog orbital visualization and analytics environment for prototyping SDA capability — orbital mechanics, sensor context, and analyst decision support in one operational picture.
| Area | Focus |
|---|---|
| Orbital analytics | Full-catalog propagation and 3D visualization; launch, reentry, and orbit-lifecycle analysis |
| Conjunction assessment | Screening, probability-of-collision analysis, risk triage, and operator decision support |
| Anomaly & maneuver analysis | Satellite anomaly detection from orbital-element regression; maneuver screening; pattern-of-life and orbital-intent assessment |
| Object characterization | Space-object characterization and cross-catalog / uncorrelated-object analysis |
| Orbital-data enrichment | Augmenting two-line element sets into higher-fidelity state estimates ("super TLEs") to improve conjunction-assessment accuracy and covariance realism |
| Space weather | Solar and geomagnetic activity as an operational input to orbital safety and launch decisioning |
| Cislunar & deep space | Custody, navigation, and awareness research beyond GEO, where publicly releasable |
- Two-Line Element (TLE) Smoothing for Increased Accuracy in Covariance Calculation — U.S. Space Force, Unified Data Library presentation, 2022
- Reverse-Ephemeris Extra-Terram Navigation for Space Exploration — DARPA position summary on route guidance, 2024
- Augmenting GEOINT with Synthetic Aperture Radar and Hyper-Spectral Analysis for Underwater Artifact Identification — NRO–NSA GEMCUTTER summary request, 2024
- Hyper-Spectral Imagery Enhancement with SAR Derivatives for Target Identification — DIU Phase II SSA summary, 2024
- Satellite anomaly-detection research recognized in DIU portfolio contests for space superiority (2021, 2023, 2025)
Related research applies synthetic aperture radar (SAR) and hyperspectral analysis to geospatial-intelligence problems (see the papers above). Open reference datasets are mirrored on this account for that work — credit belongs to the original authors; see each repository's upstream:
- LS-SSDD-v1.0-OPEN — large-scale Sentinel-1 SAR ship-detection dataset (mirror)
- SAR-2-CAE_Ship_Dataset — SAR ship-detection dataset for small-object detection (mirror)
- firestorm-satellite-data — a small production example of the orbital layer: a daily TLE pipeline for fire-relevant polar-orbiting satellites, consumed live by FIRESTORM
- keeptrack.space — fork of the open-source KeepTrack astrodynamics project, maintained as an upstream reference
Production source code, data-source integrations, government deployments, infrastructure, and non-public operational capability. Where a capability is described here, it is described at the level already public in the linked material.
DEEPSIGHT is part of the public portfolio of Deepinder Uppal — Chief Technology Officer working across artificial intelligence, space systems, and public-safety technology. GitHub profile · LinkedIn