It is difficult for a space robot to perform autonomous relative navigation of a non-cooperative space target using only a single line-of-sight measurement. To solve this problem, a decentralized-centralized relative navigation method based on multiple space robots in a leader-follower formation is proposed. All the leader and follower robots observe the same space target and combine their estimates to obtain an improved estimate of the target motion. The relative navigation filter of each leader and follower robot is independently implemented based on non-dimensional invariant sets of Hill-Clohessy-Wiltshire (HCW) equations. The invariant sets of relative motion between the follower and leader space robots are known due to their mutual cooperation, and are used in this research as state equality constraints to improve the estimate of the target motion. Numerical simulations show the feasibility of the proposed method, and the results indicate that the constrained state estimation accuracy of the space target is improved compared to unconstrained state estimation.
展开▼
