Research On Satellite Formation Relative Navigation Filter Algorithm Based On Short Arc Measurement

With the development of satellite technology,people's sights have turned from nearground space to deep space exploration.Satellite formations have become a hot spot for deep space exploration,because of its low cost,high reliability,and ability to form a large baseline.The relative navigation technology is one of its key technologies and it is the premise of achieving the mission of the formation.Considering that the ground information cannot be used for real-time navigation during the flight of deep-space exploration formations,the relative navigation method based on inter-satellite measurement information is studied in this paper.Considering the conditional constraints,the relative state estimation is achieved only by finite time measurement information.On the basis of it,through the improvement of the filtering algorithm and the addition of multiple measurement information,the performance of various aspects of the navigation system is improved.The main contents of the thesis are as follows:For large eccentricity elliptical orbit formations,linear motion models and nonlinear motion models are derived.The principle of minimum variance estimation and Bayesian estimation is studied.And Kalman filtering and extended Kalman filtering algorithm based on the two principles are given.The observation model of distance and angle measurement was established.By analyzing the noise characteristics of the model,the selection of the optimal measurement arc segment was achieved,and the idea that the noise characteristic distribution influences the estimation accuracy distribution was verified through simulation.Considering that the relative motion model is nonlinear,the CKF(Cubature Kalman filter)algorithm is studied to improve the estimation accuracy of the filter.The factors that affect the accuracy of the EKF(Extended Kalman filter)estimation are analyzed at the same time,therefore a simplified CKF algorithm is proposed.The simulation results show that for this relative navigation system,the estimation accuracy is improved and the calculation amount is moderate.Considering that the noise interference makes the model mismatch,the robustness of the filtering needs to be improved.A strong tracking filter algorithm(STF)is analyzed for this study.Through the orthogonality principle,the single fading factor and multiple fading factor algorithms are obtained.And extended it to the CKF framework,the strong tracking CKF algorithm is obtained.With the same principle,the simplified CKF algorithm proposed in the previous section is improved,then a simplified strong tracking CKF algorithm is obtained.Simulation results show that the SSTCKF algorithm is feasible and has the best robustness.In order to improve the fault-tolerance of the navigation system,so that it can still realize the navigation function in the case of some sensor failures,so added the AFF(Autonomous Formation Flyer)measurement system.An observation model is established for the measurement system,and the relative navigation algorithm based on AFF information is realized by transforming the measurement information.The federated filtering algorithm is studied to fuse the distance and angle information with the AFF information.The simulation result shows that the method is feasible;In the federated filtering,a fault diagnosis processing system was added to detect and eliminate fault information in real time,to prevent pollution of fault information,and to improve the fault tolerance of the navigation system,the feasibility was verified by simulation at last.