SMCTE - Space Traffic
A Python & TensorFlow/Keras system using LSTM and GRU neural networks to predict satellite collision probabilities from orbital CDM time-series data.
SMCTE — Space Traffic Monitoring System
Tagline: "Predicting space collisions with Deep Learning, reducing false alarms in satellite operations."
What Is It?
SMCTE is an intelligent Conjunction Assessment system developed to tackle the growing problem of Space Debris.
The Problem
With thousands of satellites in orbit, the risk of collision is real. Traditional methods based solely on orbital physics generate too many "false positives", forcing operators to waste fuel on unnecessary avoidance maneuvers.
My Solution
I developed an application that uses Deep Learning (Recurrent Neural Networks — RNNs) to analyze the historical trajectory error data of space objects. The system:
- Ingests orbital data messages (CDMs).
- Processes time-series data of position and velocity.
- Estimates the Probability of Collision with greater accuracy than static models.
Tech Stack
- Python & Data Science: Pandas, NumPy, Scikit-Learn.
- Deep Learning: Keras/TensorFlow (LSTM and GRU models).
- Interface: Streamlit (interactive dashboard for operators).