Threat assessment of non-cooperative satellites in interception scenarios: A transfer window perspective

This paper proposes a threat assessment framework for non-cooperative satellites by analyzing their motion characteristics, developing a quantitative evaluation methodology, and demonstrating its effectiveness via representative scenarios with neural network acceleration. The framework first establishes a threat evaluation model that integrates three core parameters: capability, opportunity, and hidden values. Subsequently, this research systematically investigates the critical role of transfer windows in threat quantification and introduces a transfer window-based threat assessment approach. The proposed methodology is validated through multiple representative scenarios, with simulation results demonstrating superior performance compared to conventional methods relying solely on optimal transfer windows or minimum distance metrics, enabling more nuanced threat ranking in scenarios where traditional techniques prove inadequate. To address computational demands, a neural network-based approximation system is implemented to achieve a 25,200 × speedup (0.005 s vs. baseline 126 s per 1000-sample batch) through parallel processing, maintaining 99.3% accuracy. Finally, the study explores the framework's extensibility to diverse NCS objectives. It identifies discrepancies between intention inference models and threat evaluation paradigms, providing methodological insights for next-generation space domain awareness systems.