中文摘要：

针对未来深空探测软着陆高精度实时导航的需求,提出了一种图像辅助的自主导航方案。首先通过下降段图像与落点区域地形匹配,获得着陆器相对于落点的位置和姿态;然后基于误差Kalman模型估计着陆器的状态误差,修正惯性导航的结果;在图像信息不可用的情况下,只进行惯性递推导航;这种方案既提高了导航的精度,也能保证实时性的要求。数值仿真验证了该方案的可行性,该方案对未来实际工程任务具有很大的参考价值。

英文摘要：

Future deep space exploration missions will require accurate navigation information for soft landing. This paper presents an innovative image-based navigation scheme for pinpoint landing. Firstly, the navigation extracts correspondences between the descent images and the Digital Elevation Maps of touchdown area,and then the relative positions and attitudes of the lander can be calculated. An extended Kalman filter of error model loosely integrates image information and INS. The filter computes the accurate estimation of the state error to correct the cumulative biases of INS.The filter only updates the INS equations when the image information is unavailable. This navigation system can improve the accuracy and meet the real-time requirements. Numerical simulations have demonstrated that the designed scheme is reasonable. It can be applied to future pinpoint landing missions for space exploration.