Translational and rotational motion planning for spacecraft close proximity using sampling-based methods

Abstract

For autonomous spacecraft close proximity under environments containing multiple obstacles and complicated constraints, incrementally rapid planning approaches stemming from sampling-based methods are investigated in this paper. Exploring planners are separately developed for the impulsive maneuvered translation and the piecewise constant controlled rotation, which, however, is constrained by the pointing limits coupling with relative positions during the proximity. Using a cost-informed parent-connecting strategy originating from dynamic programming as well as a sweeping growth fashion balanced between tree-based and graph-based methods, an asymptotically optimal unidirectional exploration method is proposed to search energy-efficient translational trajectory without collision. As for the rotation planning, the pointing constraints are taken as virtual obstacles in the state-space augmented with time horizon planned by the translation and, accordingly, a bidirectional exploration method is developed to generate constraint-satisfied slew paths with fast convergence rate. Numerical experiments indicate that the proposed sampling-based methods can rapidly return asymptotic optimal translation trajectory and rotation path satisfying collision avoidance and sensor field-of-view constraints.