A simulated flight software pipeline for satellite telemetry FDIR. This purely synthetic environment utilizes a Kalman-Neural Centaur architecture: Linear Kalman Filters handle deterministic physics while GPT-5.1 generates SITREPs during simulated EW jamming events. Built with Poliastro to validate theoretical FDIR logic against synthetic datasets.
Simulated Telemetry FDIR and Synthetic Electronic Warfare Detection Pipeline
Real output — physics_engine.py generates the synthetic LEO telemetry with a +5 km coordinate spoof injected at T+30s, and flight_software.py's Kalman FDIR flags and rejects it instantly (residual > 1 km threshold). The estimate then dead-reckons; the slow drift as a sustained spoof re-captures a position-only filter is a documented limitation. Regenerate with python make_hero.py.
Sentinel Node is a theoretical flight software framework developed to simulate satellite telemetry monitoring and Fault Detection, Isolation, and Recovery (FDIR) protocols within a contested synthetic environment. The project serves as a proof-of-concept for a hybrid Kalman-Neural Centaur architecture, designed to identify non-ballistic anomalies in simulated trajectory data.
This project is a purely computational simulation intended for research and educational purposes. It utilizes synthetic datasets to model how autonomous systems might identify coordinate spoofing or sensor jamming in a laboratory setting. All telemetry and threat events are generated programmatically and do not represent actual flight data or real-world satellite operations.
- Synthetic State Estimation: Employs a Linear Kalman Filter (LKF) to propagate simulated state vectors. This establishes a mathematical baseline for expected orbital positioning within the software's coordinate system.
- Simulated FDIR Logic: Implements automated isolation protocols based on synthetic thresholds. If a programmatically generated sensor measurement exceeds a 1.0 km residual, the system isolates the data and relies on dead-reckoning to preserve the integrity of the simulated baseline.
- Neural Supervision Layer: Integrates GPT-5.1 to analyze telemetry events that trigger simulation flags. The model generates technical Situation Reports (SITREPs) based on the mathematical discrepancy between the predicted state and the injected threat data.
- Orbital Simulation Engine: Generates 3D orbital ground truths using the poliastro and astropy libraries. These tools are used to create a consistent, theoretical environment for testing autonomous logic against curved orbital paths.
- Software Engineering: Python, NumPy, Linux-compatible ETL pipelines.
- Simulation Framework: poliastro, astropy.
- AI Integration: OpenAI API (GPT-5.1) using structured JSON output formatting.
- Visualization Interface: Simulated telemetry broadcasting via WebSockets to a local OpenSpace instance.
The FDIR pipeline was evaluated against a range of simulated threat intensities to determine the effectiveness of the theoretical isolation logic.
| Simulated Threat Intensity | Detection Latency | Accuracy (RMSE) |
|---|---|---|
| 0.5 km (Sub-threshold) | FAILED | 0.5046 km |
| 5.0 km (Overt Spoofing) | 0s | 0.1202 km |
| 50.0 km (Brute Force) | 0s | 0.0341 km |
These results demonstrate the system's ability to maintain a noise-tolerance envelope for minor synthetic variances while achieving instantaneous isolation for major kinematic violations within the simulation.
To maintain operational security (OPSEC) and prevent the exposure of credentials, this project utilizes environment variables for all API interactions.
- Environment Setup:
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt
- Credential Management:
Configure a local
.envfile in the root directory. This file is included in the.gitignoreto ensure keys remain local to the user's environment.
OPENAI_API_KEY=your_theoretical_access_key
- Execution Pipeline:
The user must initialize the kinematic surrogate generator to produce a synthetic
flight_telemetry.csvbefore launching the FDIR monitoring loop.
Michael Fowler is a Department of War Cost Estimator and systems analyst focusing on autonomous systems and tactical defense technology. He holds an Active Secret Clearance and is a Master of Science candidate in Systems Analysis at the Naval Postgraduate School.