挠性航天器动力学状态的自主识别与监测  

Autonomous Identification and Monitoring of Dynamic States of Flexible Spacecraft

作　　者：程天明 夏斌 魏承[1] 徐赫屿[2] CHENG Tianming;XIA Bin;WEI Cheng;XU Heyu(School of Astronautics,Harbin Institute of Technology,Harbin 150001,China;Beijing Institute of Control Engineering,Beijing 100190,China)

机构地区：[1]哈尔滨工业大学航天学院,哈尔滨150001 [2]北京控制工程研究所,北京100190

出　　处：《宇航学报》2025年第2期282-292,共11页Journal of Astronautics

基　　金：国家重点研发计划(2021YFB1715000);国家自然科学基金(62022013)。

摘　　要：卫星在轨运行时,柔性太阳翼的振动会影响其姿态稳定性并降低控制精度。为监测并优化卫星在轨状态,需准确识别太阳翼在轨振动信息。当前测量方法多依赖太阳翼表面的合作标识点,但这会影响太阳翼动力学参数,且不利于在轨操作。为此,提出了新的无标识点视觉测量和模态辨识方法。将对级约束引入特征点的识别与提取中,增强了场景特征性;设计了用于二值化图像边界特征提取的卷积核,改进了相对位置关系描述子,进而计算特征点的空间三维坐标,实现了准确、稳定的识别与监测。为验证无标识点视觉算法的精度,提出了基于自研多体动力学仿真软件MBDyn的精度评价途径。基于该评价途径开展大挠性太阳翼动力学仿真研究,得到视觉测量算法平均误差为0.214 mm。最后,搭建了柔性太阳翼地面试验平台,将地面试验平台的视觉测量信息输入到模态辨识系统中,计算太阳翼模态参数信息,得到模态辨识误差为2.63%,仿真建模误差为5%,验证了方法的有效性,为航天器在轨状态监测与航天器智能运维奠定了基础。When the satellite is in orbit,the vibration of its flexible solar wing impacts the satellite′s attitude stability and reduces control accuracy.To monitor and optimize the satellite’s on-orbit operational state,it is crucial to identify the vibration information of the solar wing.Currently,most measurement methods rely on cooperative marker points placed on the surface of the solar wing,which can alter the dynamic parameters of the wing and are not practical for on-orbit applications.The non-contact,marker-less visual measurement method offers advantages in terms of accuracy and convenience,but the absence of cooperative marker points significantly increases the measurement difficulty.To address this,a new marker-less vision measurement and modal identification method is proposed for large space flexible structures like the solar wing.By incorporating the epipolar constraint into the description of feature points,the scene′s characteristics are enhanced.A convolution kernel is designed for feature extraction of binary image boundaries to optimize the relative position relationship descriptor(RPD).Subsequently,the spatial coordinates of the feature points are calculated,achieving accurate and stable identification and monitoring.To assess the accuracy of the visual algorithm,an evaluation approach based on the self-developed multi-body dynamics simulation software MBDyn is introduced.Using this approach,dynamic simulations of the large flexible solar wing are conducted,revealing an average error of 0.214 mm for the visual measurement algorithm.Furthermore,a ground test platform for the flexible solar wing is established.The visual measurement data from this platform is input into the modal identification system to calculate the modal parameter information of the solar wing.The modal identification error is 2.63%,with a simulation modeling error of 5%.These results validate the effectiveness of the proposed method and provide a foundation for on-orbit condition monitoring and intelligent operation and mainten

关 键 词：空间大挠性结构 非接触测量 视觉精度评价 模态辨识 

分 类 号：V414.1[航空宇航科学与技术—航空宇航推进理论与工程]