A Self-Supervised Framework for Space Object Behaviour Characterisation

2025-04-08

AI Summary

To address escalating space safety challenges, this work introduces the first self-supervised foundation model for space object behavioral analysis. Leveraging 227,000 ground-based light curves, the model employs a Perceiver-VAE architecture trained via self-supervised reconstruction and masked reconstruction pretraining, followed by fine-tuning using dual physics-based simulators (CASSANDRA and GRIAL) and CAD-model-driven physical consistency regularization. The unified framework supports anomaly detection, motion pattern classification (e.g., sun-pointing, spin), and high-fidelity light curve generation. Pretraining achieves a reconstruction error of only 0.01%. After fine-tuning, the model attains an AUC of 0.90 (88% accuracy) for anomaly detection and 0.95 (82% accuracy) for motion pattern classification. It successfully identifies interpretable anomalies—such as satellite glints—in real-world data.

Technology Category

• Computer Vision: Large Vision Models
• Machine Learning: Calibration & Uncertainty Quantification
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Foundation Models, pre-trained on large unlabelled datasets before task-specific fine-tuning, are increasingly being applied to specialised domains. Recent examples include ClimaX for climate and Clay for satellite Earth observation, but a Foundation Model for Space Object Behavioural Analysis has not yet been developed. As orbital populations grow, automated methods for characterising space object behaviour are crucial for space safety. We present a Space Safety and Sustainability Foundation Model focusing on space object behavioural analysis using light curves (LCs). We implemented a Perceiver-Variational Autoencoder (VAE) architecture, pre-trained with self-supervised reconstruction and masked reconstruction on 227,000 LCs from the MMT-9 observatory. The VAE enables anomaly detection, motion prediction, and LC generation. We fine-tuned the model for anomaly detection&motion prediction using two independent LC simulators (CASSANDRA and GRIAL respectively), using CAD models of boxwing, Sentinel-3, SMOS, and Starlink platforms. Our pre-trained model achieved a reconstruction error of 0.01%, identifying potentially anomalous light curves through reconstruction difficulty. After fine-tuning, the model scored 88% and 82% accuracy, with 0.90 and 0.95 ROC AUC scores respectively in both anomaly detection and motion mode prediction (sun-pointing, spin, etc.). Analysis of high-confidence anomaly predictions on real data revealed distinct patterns including characteristic object profiles and satellite glinting. Here, we demonstrate how self-supervised learning can simultaneously enable anomaly detection, motion prediction, and synthetic data generation from rich representations learned in pre-training. Our work therefore supports space safety and sustainability through automated monitoring and simulation capabilities.

Problem

• Develops a Foundation Model for Space Object Behavioural Analysis.
• Enables anomaly detection and motion prediction using light curves.
• Supports space safety via automated monitoring and simulation.

Innovation

• Self-supervised learning for space object analysis
• Perceiver-VAE architecture for anomaly detection
• Fine-tuning with simulated light curve data