An orbital traffic simulation platform for studying satellite propagation, conjunction detection, and autonomous collision avoidance.
Themis is a space systems project focused on the growing challenge of orbital congestion.
Modern Earth orbit contains thousands of active satellites and an increasing number of potential conjunction events. As launch rates continue to rise, future space traffic management systems will need to coordinate large numbers of spacecraft efficiently and safely.
The long-term goal of Themis is to provide a simulation environment for:
- Orbit propagation
- Conjunction detection
- Collision risk assessment
- Autonomous maneuver planning
- Multi-agent satellite coordination
- Space traffic management research
The project is being developed incrementally, beginning with accurate orbital propagation using real-world satellite data.
Themis can currently:
- Load real satellite TLE data
- Parse TLEs into propagatable satellite objects
- Propagate satellite positions using SGP4
- Generate future ECI position vectors
- Produce position tables across configurable time horizons
- Calculate pairwise satellite distances
- Detect close approaches below configurable thresholds
- Scan orbital trajectories across multiple timestamps
- Export conjunction events to CSV
- Track closest observed approaches between satellite pairs
Current demonstration:
- 25 active satellites
- 24 hour propagation horizon
- 30 minute timestep
- 1000 km conjunction threshold
Output:
- Conjunction event reports
- Closest approach statistics
- CSV exports for downstream analysis
TLE Data
↓
Satellite Objects
↓
SGP4 Propagation
↓
ECI Position Vectors
↓
Position Tables
Space traffic management is becoming increasingly difficult as orbital density rises.
Before a collision can be predicted or avoided, a system must answer a simpler question:
Where will every satellite be?
Themis begins by solving that problem and gradually expands toward large-scale autonomous coordination and collision avoidance.
Current architecture:
CelesTrak TLE
│
▼
Propagation Engine
│
▼
Position Tables
│
▼
Conjunction Detector
│
▼
Event Reports
│
▼
CSV Export
Planned architecture:
TLE Data
│
▼
Propagation Engine
│
▼
Conjunction Detection
│
▼
Risk Assessment
│
▼
Maneuver Planning
│
▼
Satellite Agents
│
▼
Coordination Engine
│
▼
Simulation Dashboard
Current:
- Python
- Skyfield
- SGP4
- NumPy
- Pandas
Planned:
- Plotly
- FastAPI
- PostgreSQL
- Agent frameworks
- Scientific Python ecosystem
- Load TLE data
- Create satellite objects
- Propagate positions using SGP4
- Generate future position tables
- Pairwise distance calculations
- Close approach identification
- Configurable warning thresholds
- Conjunction event reporting
- Risk scoring
- Orbital density studies
- Statistical conjunction analysis
- Satellite agents
- Maneuver negotiation
- Conflict resolution strategies
- Thousands of satellites
- Performance benchmarking
- Coordination experiments
Themis is intended to investigate questions such as:
- How does conjunction frequency change as orbital density increases?
- Can decentralized coordination outperform centralized control?
- How many maneuvers can be avoided through negotiation?
- What tradeoffs exist between fuel usage and collision risk?
- How well do different coordination strategies scale?
Themis is still in active development.
Not yet implemented:
- Maneuver planning
- Agent coordination
- Visualization dashboard
Current propagation accuracy is dependent on publicly available TLE data and standard SGP4 modeling assumptions.
from themis.propagator import get_position
position = get_position(satellite)
print(position)Example output:
{
"satellite": 'CALSPHERE1'
'time': '2026-06-03T22:59:26Z'
"x_km": -5231.8,
"y_km": 1244.3,
"z_km": 4087.6
}Themis is being developed as an engineering and research project rather than a simple software application.
The project emphasizes:
- Measurable experiments
- System architecture
- Scientific reproducibility
- Technical documentation
- Performance evaluation
Hem Vadgama
Computer Engineering Penn State University (Schreyer Honors College)
Building systems at the intersection of space, simulation, autonomy, and software engineering.