LEOPARDD
Predicting debris before it becomes a threat
LEOPARDD (Low Earth Orbit Prediction and Analysis of Radar-Detected Debris) is an AI-powered tool developed by ASTREO to forecast the evolution of space debris in low Earth orbit.
It processes sequences of radar data to generate 3D time-dependent density maps, helping mission planners and operators anticipate high-risk regions and optimize mitigation strategies.
How it works
LEOPARDD uses a hybrid AI architecture, combining:
- ConvLSTM3D: Captures spatiotemporal patterns in sequences of radar snapshots, enabling next-frame prediction in 3D grids.
- Recursive Neural Network (RNN): Models the sequential logic and temporal evolution of debris populations over longer timescales, improving accuracy in dynamic environments.
Together, these models learn to recognize debris trends, cluster behaviors, and potential cascade effects.
