Research On The Distributed Filtering Algorithm Based On Spacecraft Relative Navigation

With the increasing development of satellite application requirements,more and more flight missions cannot be completed by a single satellite,but tend to develop in the direction of satellite formation,such as in-orbit acquisition and maintenance of faulty spacecraft,space garbage removal,constellation location acquisition and maintenance and other tasks.High precision relative orbit determination of in-orbit satellite is also a difficulty in engineering application.Nonlinear filtering theory is the key technology to solve these difficulties and the important guarantee for the success of various space missions.Satellites are developing towards modularization and miniaturization.Each module can have an independent sensor.Due to the limitations of individual sensor's own capacity and processing capacity,multiple sensor networks have strong robustness and good self-adaptability.In view of this development trend,the specific research contents of this subject are as follows:Considering the relative motion,Square root cubature Kalman filter(SCKF)algorithm is used to estimate the relative motion of the deputy star from the chief star,and it is compared with the unscented Kalman filter(UKF)to verify the superiority of SCKF filter in relative orbit estimation.Based on the relative kinematics equation and the relative attitude equation based on Rodrigo's parameter,the satellite relative motion coupling dynamic model was established.In the case of the relative position and attitude coupling,the SCKF filtering algorithm was combined to realize the coupling estimation of the relative orbit and attitude between satellites,and the results were compared with those in the uncoupled state.Based on information filtering(IF)and CI algorithm,a multi-sensor Distributed hybrid information filter(DHIF)algorithm was established to estimate the relative orbit between satellites.Sensor nodes transmit their own measurement information and receive measurement information from neighboring sensor nodes.Combined with CI algorithm,the estimation state of sensor nodes tends to be consistent.Compared with centralized IF algorithm and single sensor IF algorithm,the relative position error is estimated to verify the applicability of DHIF algorithm.For the formation flight mission of two satellites,SCKF algorithm was improved,information matrix and information state vector in the information filtering framework were introduced,and Square root cubature H-infinity information filtering(SCHIF)algorithm was established.The algorithm is based on distributed filtering structure and consistency theory.Compared with the data generated by centralized filtering and single sensor filtering,the superiority of SCHIF filtering algorithm is verified.An algorithm based on Distributed hybrid consensus-based square-root cubture quadrature information filtering(HC-SRFCQIF)is proposed.The fifth order spherical simplex radial quadrature rule is used to deal with the nonlinear system,and a new measurement update strategy is added in the hybrid consistency filter framework,and the predicted measurement error is considered,so that the algorithm can produce more accurate estimation.And the distributed filter was applied to multiple satellites formation in the navigation problems in the future war.To establish a model for our multi-formation satellites to jointly determine the absolute orbit information of enemy satellites through information exchange,each satellite as a node,a single satellite damage does not affect the ability of the whole satellite platform.