Operating Mode Identification of Satellite Solar Arrays Based on the Fusion of Temporal and Spatial Features
编号:2
访问权限:仅限参会人
更新:2025-11-10 10:12:56 浏览:30次
口头报告
摘要
The solar array (SA) is a vital component of satellite electrical power systems, providing a continuous energy supply for sustained on-orbit operations. Accurate identification of its operating modes, supported by multi-source sensing data, is essential for efficient satellite management. This study introduces a dual-branch spatial-temporal coding framework to address the limited recognition accuracy and weak dynamic perception of conventional data-driven methods. The key contribution lies in extracting complementary spatial and temporal features from sensing data and integrating them through a soft gating mechanism that adaptively fuses cross-modal information. To enhance reliability, temperature-scaled posterior probabilities are applied to calibrate model confidence, particularly during rapid transitions between dynamic operating modes. Experiments with on-orbit telemetry data show that the method achieves 99.49% accuracy, while the soft gating fusion reduces the expected calibration error by more than 21.4%. By combining multidimensional feature modeling, feature fusion, and confidence calibration, this study provides a robust and reliable solution for satellite SA operating mode identification.
关键词
satellite solar arrays,operating mode identification,sensing data,temporal and spatial features,feature fusion
稿件作者
Meng Wang
Harbin Institute of Technology
Yingting Fan
Shanghai Institute of Satellite Engineering
Guangquan Zhao
Harbin Institute of Technology
Yang Wu
Shanghai Institute of Satellite Engineering
Xiyuan Peng
Harbin Institute of Technology
发表评论