Research On Lander Position And Attitude Determination Based On Crater Fitting Ellipse For Precision Landing

As one of the key technologies in the deep space exploration, the level of autonomous navigation system impacts on the mission’s success along with the continuous development of deep space exploration mission. The precision of navigation could be boosted largely when craters are used as navigational landmarks their distinct features. With the supports of 973 Program ―Research on GNC for Precise Landing on Planets‖ and National Natural Science Foundation of China ―Tracking Strategy and Error Analysis of Independent Feature Point in Star Landing Process‖, this dissertation studies the autonomous optical navigation methods based on crater matching of deep space exploration descent and landing phase. The main contents of this dissertation are as follows:Firstly, according to the characteristic of autonomous optical navigation based on craters, the crater matching algorithm is analysed briefly. With projection relation of perspective projection considered, navigation camera model is built.Secondly, the absolute optical navigation method based on crater matching is studied by combination of high resolution craters location information from orbiter. The geometric and algebraic constraints between the marginal elliptic curves of craters and its 2D images are derived, and then the linear equations about lander’s motion are established by using Kronecker product. Consequently, the laner’s attitude and position relative to targeted features are uniquely acquired. Under the condition that the base map is accurate, this paper analyzes the relationships between the lander’s motion errors and the crater’s image noise.Then this paper proposes a landing relative navigation algorithm based on inter-frame homography in order to overcome crater matching difficulties in absolute optical navigation. Lander’s motion model in probe fixed coordinate frame is built by marginal elliptic curves correspondence of craters through inter-frame images. A new normalized homography estimation method is proposed to reduce the condition number of the measurement matrix, and enhances the accuracy of the estimated homography matrix. Combining the distance between probe and surface that is obtained by laser altimeter, the motion of probe is determined.At last, the multi-measurements fusion based navigation is researched. For the large error of relative navigation in landing fixed coordinates and the crater landmark may can not matched in absolute navigation, this paper integrates the measurements of absolute navigation and relative measurements in image sequence to capture the advantages of both measurements and greatly improve the navigation precision. At last, lander’s position error remains less than 10 m, and lander’s attitude error smaller within 0.5°.