Autonomous image-based navigation using vector code correlation algorithm for distant small body exploration
Autonomous image-based navigation using vector code correlation algorithm for distant small body exploration
复制标题
DOI：
10.1016/j.actaastro.2020.10.013
复制
发表时间：
2020-10
影响因子：
3.5
通讯作者：
Genki Ohira;Shuya Kashioka;Y. Takao;T. Iyota;Y. Tsuda
中科院分区：
文献类型：
--
作者：
Genki Ohira;Shuya Kashioka;Y. Takao;T. Iyota;Y. Tsuda
提出了一种基于图像的航天器自主导航方法，用于估计远距离小天体探测中航天器的目标相对位置。主要的焦点是在高海拔的位置估计，其中目标体的轮廓可以在图像中看到。小行星探测器Hayabusa2于2019年2月以精确的精度降落在小行星Ryugu上。在这次使命中，地面操作人员在Ryugu表面上方20公里至50米处估计了小行星的相对位置。对于探索小行星主带以外的小天体，与地球通信的延迟对于反馈制导来说是不可接受的。对于较大的物体，这种情况变得更糟，因为动力学的时间常数变得更小。因此，即使在高空，也需要实时自主导航用于远距离小天体探测。为了实现高精度、实时的自主导航，提出了一种基于地形相对导航（TRN）的自主位置估计方法，该方法通过比较标称地形信息和实际地形信息来估计偏差。除了TRN，矢量码相关（VCC）算法被用于地形信息的亮度比较。该算法是一种用于模板匹配的相关性计算方法，其找到图像中的最大相关区域。利用VCC算法，可以通过适合于FPGA的XOR运算来真实的实时计算相关。与其他方法进行了比较，该方法的估计精度和处理时间进行了评估。结果表明，该方法在真实的时间内获得了与图像分辨率相当的估计精度。最后，从隼鸟2号飞行数据的评估表明，所提出的方法的估计精度和处理时间是适合于真实的使命环境。该方法将成为远距离小天体探测的关键技术。
This paper proposes an autonomous image-based navigation method for estimating the target-relative position of a spacecraft for distant small body exploration. The main focus is position estimation at high altitude where the outlines of a target body can be seen in images. The asteroid explorer Hayabusa2 touched down on the asteroid Ryugu with pin-point accuracy in February 2019. For this mission, the asteroid-relative position was estimated by ground operators from 20 km to 50 m above the surface of Ryugu. For the exploration of small bodies farther than the asteroid main belt, the delay of communication with Earth is unacceptably large for feedback guidance. This situation becomes worse for larger bodies because the time constant of the dynamics becomes smaller. Therefore, real-time autonomous navigation is required for distant small body exploration even at high altitude. To accomplish high-accuracy and real-time autonomous navigation, an autonomous position estimation method based on terrain-relative navigation (TRN) that estimates deviation by comparing nominal terrain information and actual terrain information is proposed. In addition to TRN, the vector code correlation (VCC) algorithm is used for the luminance comparison of terrain information. This algorithm is a type of correlation calculation method for template matching that finds the maximum correlated region in images. With the VCC algorithm, correlation can be calculated in real time via XOR operations suitable for FPGA. The estimation accuracy and processing time of the proposed method were evaluated with a comparison to those of other methods. The results show that a high estimation accuracy, similar to the image resolution, was accomplished in real time. Finally, an evaluation using flight data from Hayabusa2 shows that the estimation accuracy and processing time of the proposed method are suitable for a real mission environment. The proposed method will be a key technology for distant small body exploration.