OURE (Orbital Uncertainty & Risk Engine)

OURE is a high-performance, enterprise-grade Space Situational Awareness (SSA) platform designed for orbital risk prediction, collision avoidance optimization, fragmentation modeling, and massive fleet screening.

Built for mission-critical speed and mathematical rigor, OURE processes Space-Track Two-Line Elements (TLEs), NASA CDDIS data, and CCSDS Conjunction Data Messages (CDMs). It propagates uncertainty using vectorized Monte Carlo simulations and evaluates Probability of Collision ($P_c$) using Foster's algorithm on the encounter B-plane.

Key Features

• Multi-Fidelity Physics Engine: Native SGP4 propagation combined with a High Precision Orbit Propagator (HPOP) featuring J2 oblateness, Solar Radiation Pressure (SRP), and atmospheric drag perturbations.
• NASA-Grade Integration: Supports parsing NASA CDDIS CPF (Satellite Laser Ranging) files for centimeter-level accuracy and implements the NASA MSFC Jacchia analytical atmospheric model for highly accurate solar flux drag modifications.
• Collision Avoidance (SLSQP): Mathematical maneuver optimization to find minimum-fuel 3D Delta-V vectors that mitigate collision risk below safety thresholds.
• NASA Standard Breakup Model: Simulation of hypervelocity impacts and debris cloud dispersion.
• Sensor Fusion: Extended Kalman Filter (EKF) updates to simulate commercial radar tasking and covariance collapse.
• KD-Tree Fleet Screening: Distributed epoch-bucketed $O(N \log N)$ screening of entire satellite constellations against the full NORAD catalog.
• Enterprise Observability: Fully instrumented FastAPI REST API and Celery/Redis background workers, seamlessly integrated with Prometheus and Grafana for real-time physics engine throughput and risk quantification latency monitoring.
• Interactive Visualizations: 3D ECI encounter geometry and massive orbital fleet visualization using Next.js, React Three Fiber, and Three.js.

Installation & Deployment

OURE can be run locally via CLI, launched via a lightweight web interface, or deployed as a full enterprise microservice stack.

1. Global Installation (CLI & Web)

You can install OURE globally using pip (or uv):

# Install the OURE backend package and its dependencies
pip install oure

# Start the FastAPI backend
uvicorn oure.api.main:app --reload

# In a new terminal, start the Next.js 3D Frontend
cd frontend
npm install
npm run dev

2. Enterprise Stack (Docker Compose)

For production environments, OURE deploys as a fully isolated 6-service stack including the API, Background Workers, Redis Broker, Operations Dashboard, Prometheus metrics, and a Grafana observability suite.

# Clone the repository for the Enterprise Stack
git clone https://github.com/h-rishi16/oure.git
cd oure

# Start the full Enterprise Stack
docker compose up --build -d

• Operations Dashboard (Next.js 3D Globe): http://localhost:3000 (Live Demo: http://92.4.73.195)
• API Documentation: http://localhost:8000/docs
• Grafana (Observability): http://localhost:3000 (Login: admin / admin)
• Prometheus Metrics: http://localhost:8000/metrics

CLI Usage

1. Analyze a Conjunction

oure analyze --primary 25544 --secondary 43205 --look-ahead 72

2. Avoidance Maneuver Wizard

Starts an interactive guide to optimize a fuel-efficient burn:

oure avoid --primary 25544 --secondary 43205

3. Fleet Screening

Screen thousands of secondaries against a fleet of primaries in parallel:

oure analyze-fleet --primaries-file p.json --secondaries-file s.json --workers 8

4. Space Debris Fragmentation

Simulate a "What-if" collision between two objects:

oure shatter --primary 25544 --secondary 43205 --fragments 5000

Architecture & Security

OURE enforces a strict, decoupled 5-layer architecture, hardened against Resource Exhaustion (DoS) and Numerical Singularities:

1. Core: Immutable data models (StateVector, CovarianceMatrix) and Prometheus Metrics Managers.
2. Data: Caching fetchers (SpaceTrack, NOAA F10.7), NASA CDDIS CPF parsing, and strict CCSDS CDM Parser.
3. Physics: Certified SGP4, RK45 Numerical integrators, NASA MSFC Atmospheric modeling, and SRP.
4. Uncertainty: Memory-hardened Vectorized Monte Carlo ensembles (capped at 100k samples), STM generation, and EKF Sensor updates.
5. Conjunction/Risk: TCA Golden-section search, Robust Foster $P_c$ math utilizing Moore-Penrose pseudo-inverses (np.linalg.pinv) to prevent singular matrix crashes, and SLSQP maneuver optimization.

Testing & Quality

OURE maintains strict engineering standards, verified by GitHub Actions CI/CD:

• Test Coverage: 88%+ enforced via pytest-cov across 70+ test suites.
• Static Analysis: Strict mypy typing and ruff linting.
• Numerical Stability: Joseph-form covariance updates, eigenvalue-ordered risk projection with singularity protection.

uv run pytest tests/ -v --cov=oure

License

MIT License