摘要
非合作目标航天器的位姿测量一直是各类复杂航天任务中所需解决的关键难题之一。传统的姿态测量算法多采用欧拉角以及旋转矩阵的方式描述目标位姿,计算参数较多,算法形式也相对复杂。针对这个问题,通过单目视觉识别出具有矩形结构特征的目标航天器的一个矩形面,然后在马达代数框架下,根据该矩形面两条平行边所在直线方程的差值信息直接计算出目标姿态参数,最终将姿态解算问题简化为一组四元线性方程组的求解。该算法无需知道目标尺寸,计算形式简洁,实时性高,同时也能很好地保证结果的正交性。数值仿真以及地面验证实验结果表明,该算法具有较强的稳定性,测量精度能够满足测控要求,在交会近距离范围内受两航天器间相对距离的影响较小。
Accurate measurement of relative position and attitude of non-cooperative target spacecraft is one of the key problems in complex space missions. Traditional attitude measurement algorithms use Euler angles and rotation matrix to describe the attitude, which led to more computation parameters and computational complexity. A new monocular vision measurement algorithm based on motor algebra is proposed to obtain the attitude of non-cooperative target spacecraft, which has some specific rectangular features. The attitude is calculated according to the difference between the equations of two parallel edges of the rectangle under motor algebra frame. This new algorithm has a more concise form and some computational advantages compared to the traditional methods. Numerical simulation and ground test results show that the algorithm has a strong stability, and can satisfy the measurement accuracy requirements of space missions.
作者
陈伟
陈志明
王惠南
CHEN Wei CHEN Zhiming WANG Huinan(College of Astronautics, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, Jiangsu, China)
出处
《光学技术》
CAS
CSCD
北大核心
2016年第6期516-522,共7页
Optical Technique
基金
江苏省自然科学基金(BK20130816)
航空科学基金(20150852013)
关键词
马达代数
姿态测量
非合作目标
单目视觉
motor algebra
attitude determination
non-cooperative target spacecraft
monocular vision
