Space Collision Warning and Avoidance
自主预警与规避系统
Developing machine learning-based space debris collision warning systems, enabling multi-stage autonomous avoidance decision-making, enhanced by spatio-temporal attention networks for improved prediction accuracy and real-time performance.
研究方法与技术路线Methodology & Technical Approach
数据获取与处理
Data Acquisition & Processing
从公开空间目标目录（如 Space-Track）获取 TLE 轨道数据，经 SGP4 传播器处理，提取接近事件特征。
Acquiring TLE orbital data from public space object catalogs (e.g., Space-Track), processing via SGP4 propagator, and extracting conjunction event features.
特征工程与建模
Feature Engineering & Modeling
构建高维碰撞特征空间，应用时空注意力网络（STAN）捕捉轨道状态序列中的关键模式，实现碰撞概率的精准预测。
Constructing high-dimensional collision feature spaces, applying Spatio-Temporal Attention Networks (STAN) to capture key patterns in orbital state sequences for precise collision probability prediction.
决策与验证
Decision & Validation
基于多级避碰决策框架，输出最优规避策略。通过高保真仿真和历史事件回溯测试，验证模型的有效性和鲁棒性。
Outputting optimal avoidance strategies based on a multi-stage collision avoidance decision framework. Validating model effectiveness and robustness through high-fidelity simulation and historical event back-testing.
关键成果Key Results
-
2
已发表论文
Published Papers
已发表 2 / 审稿中 1 2 published / 1 under review含 Submitted 论文
incl. submitted paper
-
1
进行中论文
Papers in Progress
含规划中论文
incl. planned papers
查看列表View list -
80%
实验进度
Experiment Progress
-
3/3验证阶段Validation
研究阶段
Research Phase
Related Papers相关论文 (2)
Learning-based framework for multistage space debris collision warning and avoidance
Liwen Yang, Jikai Wang, Ming Xu, Xue Bai
Proposes a learning-based framework for multi-stage space debris collision warning and avoidance, integrating orbit prediction, collision risk assessment, and optimal avoidance strategies.
提出一种基于学习的多阶段空间碎片碰撞预警与规避框架，整合了轨道预测、碰撞风险评估和最优规避策略。
@article{yang2026learning,
title={Learning-based framework for multistage space debris collision warning and avoidance},
author = {Yang, Liwen and Wang, Jikai and Xu, Ming and Bai, Xue},
journal={Aerospace Science and Technology},
pages={112305},
year={2026},
publisher={Elsevier}
}
Low-orbit Space Debris Warning and Autonomous Collision Avoidance for Space Environment Governance
Liwen Yang, Jikai Wang, Jun Jiang, Xue Bai, Ming Xu
Addressing the need for LEO space debris environment governance, this work proposes debris warning and autonomous collision avoidance methods, providing technical support for sustainable space environment utilization.
针对低轨空间碎片环境治理需求，提出碎片预警与自主碰撞规避方法，为空间环境可持续利用提供技术支撑。
@inproceedings{yang2025low,
title={Low-orbit Space Debris Warning and Autonomous Collision Avoidance for Space Environment Governance},
author = {Yang, Liwen and Wang, Jikai and Jiang, Jun and Bai, Xue and Xu, Ming},
booktitle={Journal of Physics: Conference Series},
volume={3015},
number={1},
pages={012005},
year={2025},
organization={IOP Publishing}
}
已投稿 / 预印本Submitted / Preprint (1)
-
STAN: A Spatio-Temporal Attention Network for Space Debris Multistage Collision Avoidance
Liwen Yang, Jikai Wang, Ming Xu, Xue Bai
Proposes STAN — a Spatio-Temporal Attention Network for space debris multi-stage collision avoidance, capturing spatio-temporal dependencies in debris motion to improve prediction accuracy.
提出 STAN——一种时空注意力网络，用于空间碎片多阶段碰撞规避，通过捕捉碎片运动的时空依赖关系提升预测精度。
@article{yang2026stan, title={STAN: A Spatio-Temporal Attention Network for Space Debris Multistage Collision Avoidance}, author = {Yang, Liwen and Wang, Jikai and Xu, Ming and Bai, Xue}, year={2026}, note={Submitted to ICLR 2026} }
其他研究方向Other Research Directions
Multi-Stage Space Debris Collision
2 篇论文2 papersInvestigating multi-stage collision chain reaction mechanisms of space debris in LEO, establishing kinetic models of...
Learn More了解更多Space Debris Environment Governance
1 篇论文1 paperExploring sustainable governance strategies for the space debris environment, including active debris removal...
Learn More了解更多