Study on Autonomous Landing Guidance Method for Spacecraft

深空探测自主着陆系统的研究主要集中在传感器、感知信息处理、基于图像的地形识别、航天器导航、制导与控制等方面。为了评估着陆系统的性能，开发了具有行星表面建模、感官模型生成、航天器动力学设计、导航制导等多种功能的图形化着陆模拟器。提出了一种基于运动立体视觉技术和边缘、阴影信息的全局映射方法。作者开发了一种建模方案来构建各种行星表面，包括陨石坑和岩石。提出了一种行星图像特征提取方法和一种利用距离信息和图像信息估计行星相对于地表的相对位置、速度和姿态的导航方法。作者还成功地开发了用于避障和精确导航的地形识别，并将航天器引导到着陆点。计算机数值仿真和图形化着陆模拟器验证了所提方法的正确性和有效性。这些结果表明，构建具有鲁棒性和高可靠性的自主着陆系统是可能的。良好的着陆系统有望为行星科学做出贡献。

Autonomous landing system for deep space exploration has been studied focusing on sensors, sensory information processing, terrain recognition based on image, and navigation, guidance and control of spacecraft. To evaluate the performance of the landing system, the graphical landing simulator has been developed, which has various kinds of functions, planetary surface modeling, sensory model generation, spacecraft dynamics design, navigation and guidance etc. The authors proposed a global mapping method based on moving stereo-vision technique and rim and shadow information. The authors developed a modeling scheme to build various kinds of planetary surface including craters and rocks. The authors also proposed a method to extract features on planetary images and a navigation method to estimate the relative position, velocity and attitude with respect to the surface by using range information and image information. The authors also succeeded in developing terrain recognition for obstacle avoidance and precise navigation and guidance of spacecraft to the landing site. The validity and effectiveness of the proposed methods were confirmed and verified by computer numerical simulations and the graphical landing simulator. These results show that it is possible to build good autonomous landing system which has robustness and high reliability. Good landing system is expected to contribute to the planetary science.