Speaker
Description
To achieve the autonomous landing on the asteroid surface, the landers typically rely on optical sensors such as cameras or LiDARs to get the navigation features’ observation information from the asteroid surface. This information is then used to update the position and attitude estimate results of the landers. Considering that the navigation features are unevenly distributed on the surface and the initial state estimate error is generally large, the lander is expected to approach the navigation features as closely as possible during the landing process to enrich the observation information. In order to plan such a trajectory, the energy optimal guidance which is widely used in asteroid landing missions is revisited in this paper. According to the energy optimal guidance law, the lander will follow the landing trajectory that consumes the least energy consumption. On the other hand, due to the pursuit of reducing energy consumption, it may result in a lack of navigation features that can be observed within the field of view during the landing process.
To solve this problem, a method of asteroid landing trajectory planning considering the observation requirements of navigation features is proposed. On the basis of energy optimal guidance, a new landing trajectory is designed. The motion of the lander is decomposed into the horizontal plane and the z axis. In the z-axis direction, the lander still moves according to the energy optimal guidance. In the horizontal plane, the lander derails to a certain extent from the energy optimal trajectory so that it is able to approach and observe the navigation features that would originally be outside the field of view. To this end, Bezier Curve is introduced into trajectory planning. Firstly, a K-means clustering algorithm is used to partition the navigation features on the asteroid surface. The center points of each group obtained by the algorithm are regarded as the control points of the Bezier Curve, upon which an observation-driven trajectory in the horizontal plane is generated. Then, the motion of the lander in different directions is combined to obtain the complete three dimensional landing trajectory with improved navigation observation performance. Based on the results of trajectory planning, the control sequence of the lander during the landing process can be obtained by tracking the trajectory with a piecewise energy optimal guidance. Finally, the effectiveness the proposed method is validated through an asteroid landing-based mathematical simulation.