Austin Barrow, Hyperios Technologies; Benjamin McCrossan, Hyperios Technologies; Phillip Vu, Hyperios Technologies
https://doi.org/10.64861/MAQJ5920
Keywords: Space Domain Awareness (SDA), GNSS Ionosphere, Orbital Launch Detection, Real-Time, Rocket-Induced Shockwaves, Atmospheric Sensing, Signal Processing, Total Electron Content (TEC), Automated Pipeline, Blind Detection, Acoustic-Gravity Waves (AGW)
Abstract:
The primary objective of this work is the development and validation of a robust, automated system to detect rocket launch events by identifying rocket-induced ionospheric disturbances with ground-based Global Navigation Satellite System (GNSS) receivers. Traditional Space Domain Awareness (SDA) relies heavily on ground-based radar and optical sensors, which face limitations from line-of-sight constraints, geographical gaps, and cueing requirements.
The ionosphere is a partially ionized layer of Earth’s upper atmosphere beginning at an altitude of roughly 100 km. This is subject to frequent variation due to diurnal changes, and effects from the lower atmosphere as well as solar and geomagnetic activity. The ionosphere can be measured using GNSS receivers that compare the relative delay between different frequencies which corresponds to the density of the plasma. Energetic events like earthquakes, tsunamis, and meteors as well as explosions and rocket launches produce disturbances in the ionosphere that are observable on GNSS receivers. This enables widely distributed GNSS receivers to provide early detection of orbital rocket or ballistic missile launches. This turns globally distributed GNSS receivers into a resilient passive sensing layer for SDA and missile warning.
This effort was validated against an extensive empirical dataset of orbital rocket and ballistic missile launches, paired with matching control periods of quiet ionospheric activity. The data encompasses events from Vandenberg Space Force Base and Cape Canaveral Space Force Station, as well as North Korean missile tests monitored by receiver networks in Japan and South Korea. This research demonstrates that GNSS networks can function as a cost-effective, persistent, global “atmospheric sensor” capable of identifying launch signatures even in “blind” scenarios without external cueing.
This real-time system uses a robust ingestion framework that streams 1Hz GNSS observations using BKG NTRIP Client (BNC). This continuous stream feeds into a cloud-based processing engine, converting dual-frequency GNSS observations into calibrated Total Electron Content (TEC) products with sub-millisecond execution latency. The engine leverages pseudorange and carrier phase observables for flexible leveling and debiasing routines modified specifically from traditional batch paradigms for live streaming execution. This TEC processing engine was validated for double-precision mathematical parity with comparable batch-processing pipelines. The continuous TEC stream serves as the foundation for downstream detection engines.
To capture the full spectrum of launch signatures, the architecture incorporates two distinct detection engines detecting distinct ionospheric phenomena: a Shock-Acoustic Wave (SAW) detector tuned for short-period pressure impulse waves, and an ionospheric TEC depletion detector engineered to identify large-scale plasma depletions caused by rocket exhaust plumes recombining with the free electrons.
The SAW and TEC depletion processing pipelines include single-station and multi-station layers tailored specifically to the physical characteristics of each phenomenon. Single-station links isolate signature waveforms within the TEC observations of individual satellite-receiver pairs. The SAW detection system isolates and identifies the N-shaped waves common to SAWs. The TEC depletion system detects rapid negative baseline shifts in localized plasma density caused by chemical recombination with exhaust plumes.
Multi-station coordination aggregates single-station links spatially to confirm spatial and temporal coherence across network nodes and verify a common propulsion source. To estimate the horizontal location of the disturbance, 2D projections of the satellite-receiver line-of-sight path are made at fixed altitudes. Candidate launch events from both the SAW and TEC depletion detectors are cross-correlated and back-projected along geodesic paths toward candidate launch sites to synthesize a multi-variate confidence score based on physical launch correlations. To achieve maximal detection performance and minimize space weather false alarms, an ongoing multi-modal fusion framework combines outputs from both engines into a unified decision matrix.
The system was evaluated through a two-tiered validation framework: algorithmic refinement using a synthetic simulation package with Monte Carlo injections of launch signatures across varied atmospheric noise regimes, followed by validation against the multi-region empirical launch dataset demonstrating robust sub-millisecond operational performance. Future work aims to expand the pipeline from a localized launch occurrence detection to full 3D trajectory identification, rocket state vector estimation for direct SDA orbit determination initialization, and rocket vehicle and yield/strength characterization.
This research provides the SDA community with a scalable solution for monitoring global launch activity and characterizing propulsion signatures through persistent, cost-effective atmospheric sensing, directly addressing the need for diverse and resilient space surveillance architectures.
Date of Conference: September 15-18, 2026
Track: Space Domain Awareness