High-accuracy target tracking estimation based on dual Kalman filter

For cooperative spacecraft driven by continuous force, the dual unscented Kalman filter (DUKF) was used to estimate the state and acceleration. The state filter and parameter filter can cooperate mutually to ensure higher filter accuracy of estimation for motions and parameters, so that the motion tracking of cooperative targets can be realized. Compared with cooperative targets, non-cooperative targets perform maneuvering with unknown forces and occurrence time, so this brings more challenges for information acquisition and state estimation. For non-cooperative spacecraft, we formulated the relative motion equation, and utilized the observation information from space-based platform for state estimation. Two extended Kalman filters (EKFs) and semi-latus rectum were used for maneuvering detection to estimate the motion state of maneuverings with multiple unknown pulses. The simulation results show that DUKF has faster rate of convergence and better performance of tracking error for state and acceleration estimation of cooperative spacecraft. For non-cooperative spacecraft, the effectiveness of the maneuvering detection strategy combining with the filter switching strategy is verified through comparison. The proposed method detected multiple maneuvers and reduced misjudgments.