RRTZ: a Path Planner Designed for Zero Gravity

Future spacecraft in zero-gravity proximity operations are expected to host autonomous planning and control capabilities. Available path planning approaches have high computational complexity or are not specifically designed for use in zero-gravity environments, particularly given multi-step inspection activities. This paper extends the Rapidly-exploring Random Tree (RRT) algorithm to a Rapidly-exploring Random Tree for Zero-gravity (RRTZ) formulation designed to efficiently generate low-energy paths in zero-gravity. RRTZ exploits drag-free locomotion for constant velocity path segments along with the tree expansion structure employed in RRT to generate paths with minimal change in local path heading. RRTZ, RRT* (an asymptotically optimal extension of RRT), and a modified version of RRT* are applied to several three-dimensional zero-gravity environments with obstacles of varying geometries. RRTZ produces paths of lower average cost, reduced average planning time, and greater robustness (100% task completion) compared to RRT* (71% task completion) and a modified version of RRT* (93% task completion). RRTZ offers a novel RRT-derived formulation for solution discovery in a variety of emerging space robotic planning tasks.