Reseach On Autonomous Optical Navigation Scheme In Asteroid Probe Landing Period

With the success of the U.S."discovery plan" and the Japanese MUSES-C project,asteroid exploration activities designed to revealing the solar system and the origins oflife have attracted research institutes’ attention in many countries. Asteroid explorationactivities have became an important branch of deep space exploration. The feature ofasteroid explorantion landing segment include a complex dynamic environment, a longdistance between ground control station and probe, communication delay and highprecision landing demanding. These characteristics have high requirements of guidanceand control capabilities, especially autonomous navigation. Research on autonomousnavigation method has become an important research topic in asteroid probe.In this paper, asteroid exploration mission landing segment as the researchbackground, optical navigation methods were studied. At the beginning, the model ofoptical navigation measuring and vector measurement method for determining therelative position and attitude based on the use of a combination of navigationrangefinder and cameras were introduced. At the same time, a method of extracting andtracking optical features was introduced here.The navigation filtering methods were studied here，by reason that the navigationmethod based on optical vector measurement accuaracy was not high with a strongnoise environment. After the establishing of the asteroid probe landing paragraghkinetic equation, the navigation filters based on EKF and UKF were designed. As therangefinder noise characteristics are time-varying during the descent of the probe, anavigation filters was designed using a simplified Sage-Husa adaptive filtering method.Gravitational field model and guidance law were designed as a case study of Eros433asteroid detection. Various filtering methods were tested in this model.Navigation measurement methods based on the line of sight were studied here. Themethod use nonlinear estimation algorithm and filtering algorithm to determine therelative position, velocity and attitude of the probe. For Gaussian least squaresdifferential correction algorithm (GLSDC) shortcoming of computationally intensive, aposition and orientation decoupling algorithm was designed here, based on theralationship between the distance in different coordinate systems. Simulation resultsshow that this method compared with GLSDC algorithms, while ensuring the samenavigation accuracy, can reduce the solver time and impove navigation instantaneity.