Research On The Autonomous Navigation Method For Small Celestial Landing Based On Optical Features

Autonomous navigation is one of the critical technologies of small celestial landing.The autonomous navigation of the landing device based on the surface optical features of small objects is a hot issue at present. In this thesis, the navigation scheme design and autonomous pose estimation are studied based on the small object landing. The autonomous navigation method for small celestial landing based on optical features is studied. The main contents of the paper are as follows:First, coordinate system and dynamic models of landing detector which will be used in the following navigation methods design, and two different optical observation models are established. Based on observability analysis of different measurement models at different number feature points, two different navigation schemes are designed.Second, the autonomous optical navigation filter algorithm is analyzed. To solve the problem of the nonlinear of navigation system, the Gaussian least square differential correction correction method and Levenberg Marquardt(LM) algorithm which are two kinds of nonlinear optimal estimation method are analysised respectively. At the same time,two different recursive filtering algorithms which are UKF and EKF are studied. And the performance of different algorithms is summarized.Then, based on the optical navigation scheme of the vision measurement, the autonomous pose estimation method based on the improved correction is studied. Based on the analysis of traditional filtering estimation performance, a kind of improved small celestial body landing autonomous position and pose estimation method which combines the optimal estimation algorithm and UKF algorithm is gived. This method to increase the precision of the estimation of landing navigation.At last, the autonomous pose estimation method based on vector measurement is studied. Through the analysis of model parameter uncertainty influence on navigation precision and model parameters is extended into the navigation state estimation, gives the small body landing state parameter information of joint estimation method to improve the model of uncertain state estimation robustness of landing conditions.