行星着陆动力下降段相对视觉导航方法

针对行星着陆动力下降段视觉导航自然路标匮乏的问题，提出了一种相对视觉导航方法。该方法利用相机和雷达的测量信息构建相对导航坐标系并求解随机视觉特征点在该坐标系下的位置矢量，利用求解得到的特征点为导航参考，设计相对导航系统，估计着陆器在相对导航坐标系下的位置、速度及姿态信息。同时，构建可观性矩阵，解耦分析位置和姿态的可观性。通过可观性分析可知利用相对导航坐标系下的一个随机特征点即可实现着陆器全状态可观。最后通过仿真分析着陆器状态误差，验证了可观测度理论分析及导航性能。该相对导航方法无需行星地形数据库，且可以实现着陆器全状态的高精度估计，满足行星精确软着陆的需求。

Relative Optical Navigation in Powered Descent Phase of Planetary Landings

A relative navigation method using a camera and a three-beam ranger is developed to address the challenge of scarce of natural mark for absolute navigation. The camera is used to measure the coordinates of the random features in the photo plane, while the three-beam ranger is used to obtain the normal vector of the landing plane. Combining these two kinds of the measurements, the relative navigation reference system is established and the position of each feature in this reference system is obtained. Exploiting the features with known positions, the navigation system is designed to estimate the lander’s position, velocity and attitude in the relative navigation system. Afterwards, the observability analysis is performed based on the observability matrix to reveal the least amount of features to satisfy the observability. Finally, the numerical simulations are performed to analyze the state estimation accuracy and validate the observability analysis. The proposed method needs no Martian terrain database, and realizes the high-precision full-state estimation and meets the requirements of Mars high-precision soft landing.