Study On Autonomous Navigation Methods Of Mars Approach Phase

As the key technology of Mars exploration,the autonomous navigation of Mars approach phase and its performance determines the Mars landing accuracy,further relates to mission success.One of the research focuses of Mars landing technology is investigating advanced approaching autonomous navigation and improving the state estimation accuracy.This Thesis,based on the National Basic Research Program of China(973 Program)—The Research of guidance,navigation,and control problem for Precise Planetary Landing,aims at meeting the requirement of autonomous approaching navigation in accuracy,real-time property,and autonomy,and investigates the autonomous navigation methods.The main work of the thesis is organized as follows.First,the feature of the autonomous measurements in Mars approach phase is analyzed.According to the geometric configuration of the Mars approaching trajectory,the linear dynamics of the Mars explorer and the observation models in B-plane coordinate system are developed.In order to analyze the effect of initial state uncertainty and observation errors on the navigation performance,the linear covariance method in B plane is proposed.Using this method,the performance of each observation technology in correcting the error covariance of explorer’s states is analyzed,which provides reference for navigation scheme design with consideration in the application field of different observation technologies.Next,required by the autonomy and real-time property of Mars approaching navigation,the autonomous navigation scheme based on Time of Arrival and Density of Photon Flux of X-ray pulse is developed.The signal models of X-ray pulsar and corresponding observation models are established.The measurement accuracy and state estimation accuracy are analyzed using theory of Cramér-Rao bound.Considering the limited navigation source,the performance index based on system observability is exploited to optimally select the X-ray pulsars.The optimal state estimation is obtained using a nonlinear estimation filter.Numerical simulations are performed to verify the navigation performance.Considering the characteristic of the terminal of Mars approach phase,the defect of the navigation method using only X-ray pulsar measurements are dug out through the nonlinear filter theory.Then,focusing on the terminal divergence caused by the strong nonlinearity using Xray pulsar-based navigation,the inter-satellite measurements aided X-ray pulsar navigation is developed to overcome the accuracy defect due to the long-cycle measurement of X-ray pulsar.By combing the X-ray pulsar measurements and the inter-satellite measurements,the navigation performance is further improved.In order to overcome the effect of velocity on reconstructing the pulse profile,the navigation results based on inter-satellite radio measurements are used to correct the X-ray pulsar measurements,thus,improving the measurement accuracy.Based on the information fusion theory,a decentralized navigation filter is designed to deal with the time-delay and non-synchronized measurements,and to obtain the robust,accurate state estimation.Afterwards,the ephemeris error,direction error of the X-ray pulsars,and the orbit errors of the radio navigation beacon are addressed by developing a relative observation scheme.The difference of the X-ray pulsar measurements from the Mars approaching explorer and the Mars orbiter is used to diminish the observation model errors.States of both the Mars explorer and the Mars orbiter are simultaneously estimated by a navigation filter.Besides,the navigation performance is analyzed through observability analysis,which is taken as the reference for optimizing the observation selection.Furthermore,the application of the proposed relative-observation-based navigation method is extended into Mars non-spherical dynamics,and the cooperative orbit determination by observing only one pulsar is realized.Finally,the numerical simulation system of autonomous navigation for Mars approach phase is developed.Involving multi-models,algorithms and databases,the complex environment of deep space is simulated.The simulation system has the function of,such as,approaching orbit design,generation of arrival sequence of pulsar’s photon,simulation of navigation sensor,design and optimization of observation scheme.The proposed navigation methods in this thesis are validated through the developed navigation simulation system.